Notes on Gramineae

:

if

t i
ui
fal pled

aes
Pisiags

rare
ae

New Yark
State College of Agriculture

At Cornell University
Dthaca, N. Y.

Library
Gift of

canes Tracey F. Hubbard... _.

mii

Cornell University

Library

The original of this book is in
the Cornell University Library.

There are no known copyright restrictions in
the United States on the use of the text.

http://www.archive.org/details/cu31924001687023

wt

DECEMBER 24, Price 6s.

THE JOURNAL

OF

THE LINNEAN SOCIETY.

Vou. XTX. BOTANY. Nos. 115-116.

CONTENTS.

Page
ma Qn Central-African Plants collected by Major Serpa
tf ¢ Pinto. By Prof. Count FicatHo and W: P. Hizey,
‘ " MA. FILS. (Abstract POPPING.) sas seie'vee cox voeeeeey 13 .

: Notes on Graminex. By Grorce Bentuam, F.R.S... 14—

si 2 2p DH EL TREC Scale

III. Report on the Arctic Drift Woods collected by Capt.

. ’Feilden and Mr. Hart in 1875 and 1876. By W. R.

4 M'Nas, M.D., F.LS., Professor of Botany, Roy. Coll.
2 of Science, Dublin. (With a woodcut.) ......0.000.... 135

\ LONDON:
SOLD AT THE SOCIETY’S APARTMENTS, BURLINGTON HOUSE,
AND BY
LONGMANS, G. EN, READER, AND DYER,
AND

iY

WILLIAMS AND NORGATE.
1881.

fi

i ‘ pe Te
y¥ K AG & LINNEAN SOCIETY OF LONDON.
— ANNIVERSARY MEETING, TUESDAY, MAY 247u, 1881.
674 Be

Prof. Auumanx, LL.D., F.R.S., President, in the Chair.—There was a numerous
attendance of the Fellows.—The Treasurer presented his Annual Account,
see page 3 of wrapper.—Afterwards the Secretary (Mr. B. Daydon Jackson)
read his Report. Since the last Anniversary 11 Fellows of the Society
had died and 4 had withdrawn. Against this 87 new Fellows had been
elected, besides 1 Foreign Member and 1 Associate. During the past year
there had been received as Donations to: the Library 106 volumes and 125
pamphlets and separate memoirs. From Scientific Societies in exchange
there had been received 96 volumes and 248 detached parts of publications,
besides 23 vols. of Donations from Editors of independent periodicals. Some
90 vols. had been purchased, viz. 80 separate and 68 serials equal to 10 vols.
The total additions to the Library being 815 vols. and 873 separate parts.
Framed water-colour sketches of Dr. Rob. Brown’s birthplace, bis London resi-
dence, and of Sir Joseph Banks's Library had been presented by Mr. R. Kippist.
The Society’s Collections and Herbaria had been duly examinedand reported
on to the Council as in good condition. After 50 years’ service Aft EDs
had resigned his position as Librarian to the Society, and the Counc, in ac-
knowledgment thereof, had granted him a retiring pension.—Thereafter pre-
sentation of portraits of the late Mr. John Miers and of Prof. St. George
were made.—Prof, Allman then delivered his Anniversary Address, “ Rec
Advances in our Knowledge of the Development of the Crenorpyora.”—Tixe
Secretary afterwards read Obituary Notices of the several Fellows, making\
special mention of Mr. E. R. Alston (late Zool. Secretary), Mr. John Gould’
(Ornithologist), Mr. Gerrard Krefft (of Sydney), Dr. Lauder Lindsay, and R.
A. Pryor, of Herts—The Scrutineers having examined the ballot, then re- ,
ported that Mr. A. W. Bennett, Mr. F. Darwin, Prof. E. R. Lankester, Sir J.
Lubbock, and Mr. G. J. Romanes had been elected into the Council in the room
of E. R. Alston (deceased), Dr. T. Boycott, Prof. M. Foster, Dr. J. G. Jeffreys,
and Prof. Mivart, who retired ; and for Officers, Sir J. Lubbock as President,
F. Currey as Treasurer, B. D. Jackson and G. J. Romanes as Secretaries.

*

a8 ee! 2)
2 20.48 3 LSl OF THE OFFICERS AND COUNCIL.

PRESIDENT.
Sir John Lubbock, Bart., M.P., D.C.L., LL.D., F.B.S8.

VICE-PRESIDENTS.

Prof. G. J. Allman, LL.D., F.RB.8. Charles Baron Clarke, M.A., F.G.S.
George Busk, F.R.S., F.G.8. Frank Crisp, B.A., LL.B.
SECRETARIES. ,
B. Daydon Jackson. | George J. Romanes, F.R.S.
TREASURER.
Frank Crisp, M.A., LL.B.
COUNCIL.
Prof. G. J. Allman, LL.D., F.R.S. Arthur Grote, F.G.S., R.A.S.
Alfred W. Bennett, M.A., B.Sc. B. Daydon Jackson.
George Busk, F.R.S8., F.G.S8. Sir John Kirk, K.C.M.G.
Charles Baron Clarke, M.A., F.G.S. Prof. E. Ray Lankester, F.R.S.
Frank Crisp, B.A., LL.B. Sir J. Lubbock, Bart., LL.D., F.R.S.
Rev. James M. Crombie, M.A., F.G.8. | Robert MrcLachlan, F.R.S.
William Sweetland Dallas. George J. Romanes, F.R.S.

Francis Darwin, M.B.

LIBRARIA‘.
James Murie, M.)., LL.D.

ASSISTANT IN THE LIBRARY.
James West.

ON AFRICAN PLANTS COLLECTED BY MAJOR SERPA PINTO. 13

On Central-African Plants collected by Major Serpa Pinto.
By Prof. Count Ficarno and W. P. Hiery, M.A., F.LS.

(Abstract, read June 16, 1881.]

Tue specimens herein discussed were collecled by Major Serpa
Pinto, during the month of August 1878, along the upper course of
the river Ninda, an affluent of the Zambesi, on the west side of
the high plateau. As regards the climate of this locality, the
temperature is described as variable, the weather as very dry
during seven or eight months of the year, and very wet during
two or three months. The nature of the soil is metamorphic
argillaceous schist; the latitude iy 14° 46’ S., the longitude
20° 56!’ E., and the elevation 1143 metres above the ocean.

The present little collection consists of seventy-two numbers,
comprising sixty-five species in thirty-nine genera; more than
a quarter of these species are new or not previously described
and published, and at least one new genus appears amongst
them. Some of the specimens are imperfect and have been diffi-
cult of final determination, especially the grasses and sedges; the
greater part have had their approximate position ascertained ;
five specimens are hopelessly defective, and accordingly have
been excluded from the examination. _

As in the case of the previously known species, the affinities of
many of those of the present collection are not only with the flora
of Huilla, in South Angola, but also, in several instances, with
that of extratropical South Africa; only a few of the species are
widely distributed in the tropics of this and other continents.

This paper, with illustrations, will appear in full in the Society’s
‘ Transactions.’

LINN. JOURN.-BOTANY, VOL. XIx. D

14 MR. G. BENTHAM ON GRAMINESZ.

wee otes on Graminew. By Gorge Bentuay, F.R.S.
[Read November 3, 1881.]

Gramtnex, so long believed to be the largest Order amongst
Monocotyledons, must now yield the palm to Orchidew in respect
of number of species; but they must still be acknowledged as
immensely predominant, as well in individual numbers as in the
part they take in the vegétation of the globe. The great majority
of Orchidex are very local, and amongst the few that are spread
over wider areas it is frequently only in a few individuals dotted
here and there; whilst a considerable proportion of Graminew are
almost cosmopolitan in their geographical distribution within or
without the tropics, often covering the ground with innumerable
individuals. Orchides are difficult to preserve; collectors bring
home but few specimens from their chief stations in tropical
lands, and those few often imperfect. Their study is therefore
surrounded by many impediments, and, with the exception of the
few European ones, is in the hands of very few botanists; whilst
Grasses, easily dried, abound in herbaria in specimens readily ex-
hibiting their most essential characters ; and every local botanist
considers himself perfectly competent to describe as new species or
genera suggested only by comparison with the few forms known
to him from the same limited locality. The consequence is that
amongst the large number of new species of Orchidez described
of late years the great majority (always excepting garden hybrids
or varieties) appear to be really distinct; whilst the number of
bad species and genera of Gramine with which science has been
overwhelmed is truly appalling. Looking to the future, itis only
probable that the preponderance in number of species of Orchider
over Graminee is likely to be greatly increased as well by new
discoveries among the former, as by a critical revision of old
species of the latter. On the other hand, although the interest
in Orchides has been so much intensified of late years, as well by
the extent to which they are cultivated as by the singularities
observed in their fertilizing-apparatus, yet their importance in
the study of the history and development of vegetation, and in
their application to the uses of man, remains as nothing compared
to that of Graminez.

This paramount importance of the latter Order in an economical
point of view has called forth innumerable treatises, memoirs, and
essays on cereals, on forage and other cultivated grasses, on

MR. G. BENTHAM ON GRAMINER. 15

meadows and pastures, on ornamental grasses, on the physiology
and properties of the Order, &c., to which I need not further
allude, my present object being merely to consider Gramines with
reference to their classification and affinities. In a systematic
point of view, the great mistake of Linneus and the earlier syste-
matists was the attempt to regard the whole spikelet as a single
flower, with a calyx and corolla to be compared with those of the
more perfect Monocotyledons. Robert Brown, with his usual
sagacity, pointed out this and other errors, and first laid down
the true principles upon which the Order could best be divided into
tribes and genera; but he unfortunately took up the idea that
the so-called lower and upper palez represented three outer seg-
ments of a perianth; and although this theory has long since been
proved to be groundless, especially by Hugo Mohl, whose views
have been fully confirmed by all subsequent careful observers,
yet so great is the authority so deservedly attached to every
thing that has issued from the pen of Brown, that his explana-
tion of the structure of the spikelet is still allowed to influence
the terminology adopted in generic and specific descriptions.
Shortly after the publication of Brown’s ‘Prodromus,’ Gra-
mines were taken up by several French botanists who had
acquired materials, rich for the time, chiefly from North America
and the West Indies. Some of these had already been published
by Michaux or by Persoon, with more or less of assistance from
Louis Claude Richard, to whom the credit of all that is good in
Persoon’s ‘Synopsis’ as well as in Michaux’s ‘Flora’ has been
attributed by several subsequent writers. The greatest value is
justly attached to all of the elder Richard’s observations in every
Order that he worked up; and there is no doubt that such assist-
ance as he gave to those two works added much to their import-
ance; but we know that he declined to attach his name to Persoon’s
Synopsis, chiefly from an unwillingness to sanction the arrange-
ment under the Linnean system, and we are by no means assured
that there may not have been other details in both works which
he did not concur in. We therefore are not justified in fixing on
him a responsibility which he refused to undertake; and the
genera and species first published by Michaux or by Persoon
should be quoted as theirs and not Richard’s, except where
Richard’s name is expressly attached to them. Michaux’s ‘ Flora’
was published in 1803, the first volume of Persoon’s ‘Synopsis’
in 1805, both of them therefore antecedent to Brown; but two
D2

16 Mk. G. BENTHAM ON GRAMINES.

other special agrostologists, Desvaux and Palisot de Beauvois, had
ample time to avail themselves of Brown’s work. Desvaux pub-
lished his new genera ina memoir which first appeared in abstract
in the ‘ Nouveau Bulletin de la Société Philomathique’ for 1810,
and afterwards in full in the first yolume of his second ‘Journal
de Botanique’ in 1813. Between these two periods Palisot de
Beauvois published his ‘Agrostographie,’ in which he undertook
a general arrangement of the whole Order, with definitions as well
of the old-established genera as of a large number of new ones,
including those of his contemporary Desvaux. The majority of
these genera have since been adopted ; but his arrangement of
them was far too technical and his characters often so vague, that
they could in most instances scarcely have been identified, were it
not for the names of the species which he refers to them and for
the really good analytical drawings accompanying his work. As
it is, several of his names have been misapplied by subsequent
botanists, who have not paid sufficient attention to, or have not
seen, those drawings.

A few years later, three eminent botanists undertook the
general study of Graminer. Kunth at Paris and afterwards at
Berlin, Trinius in Germany and afterwards at St. Petersburg,
and Nees von Esenbeck at Bonn, afterwards at Breslau, worked
more or less contemporaneously, but with little or no communi-
cation with each other. Kunth’s ‘ Revisio Graminum,’ published
in 1829 and following years, is a work not only splendidly illus-
trated, but remarkable alike for the accuracy of detail in the descrip-
tions of species, as for several of the views given of their structure
and arrangement. This work, however, is so costly as to be acces-
sible to few botanists, and the more generally known first two
volumes of his ‘Enumeratio Plantarum,’ containing the Grasses,
were unfortunately a far too hasty compilation. He had entered
into an agreement with old Cotta for the preparation of a com-
pact Synopsis Plantarum on the plan of Persoon’s, and had
received a considerable sum of money on account of the work ;
but when it came to the actual drawing it up, Cotta insisted upon
its being arranged according to the Linnean system, which Kunth
would no more agree to than did the elder Richard in the case of
Persoon. The Synopsis or Enumeratio was therefore still in
abeyance when old Cotta died; and his successors, not caring for
the special plan adopted, insisted on an immediate return for the
money advanced ; azd I several times heard Kunth himself much

MR. G. BENTHAM ON GRAMINER, 17

bewail the necessity he was under of getting up these volumes
without the care and study he could have wished to bestow on
them, and which he did apply to his next volume on Cyperacee.
Kunth also in all his works fully adopted Brown’s theory as to
the homology of the parts of the spikelet, carrying it out in detail
to a degree which sometimes amounts almost to a reductio ad
absurdum ; as, for instance, in Piptatherum and Milium, two
genera so closely connected in structure that they are still regarded
by many experienced botanists as slightly different sections of
one genus. In both genera we see the whole spikelet consist of
two similar outer glumes without the slightest rudiment of a
flower in their axis, and of a third glume enclosing a flower and
its palea; and yet we are told that whilst in Piptatherum we
have two glumes and one flower, we must in Miliwm consider
them as one glume and two flowers.

Trinius published his ‘ Fundamenta Agrostographie’ in 1820,
something on the plan of Beauvois’s ‘ Agrostographie,’ but evi-
dently founded on insufficient materials and bibliographical re-
sources, and with some neglect of the already well-established rules
of nomenclature. From that time, however, he devoted himself
with the greatest zeal and increasing success to the study of the
Order. I heard him say, 4 propos of some rather costly collection
of specimens, that he would willingly sell his last coat for a new
grass; and all his later works, down to his last papers worked up
in conjunction with Ruprecht, and published in the Memoirs
of the Petersburg Academy, are of the greatest value to agros-
tologists, though he never followed them up by any general synop-
tical view of the Order. In respect of terminology, he so far
modified that of Kunth, that where a glume is theoretically
supposed to have a flower in its axil, but reaily has not even the
slightest rudiment, he does not, like Kunth, call it a whole
(neutral) flower, but only half a flower.

Nees von Esenbeck never confined himself so exclusively to
Graminez as did Trinius; he never published any general con-
spectus of the Order, and entered but little into general consi-
derations of their structure and terminology; but he described
with great care the grasses of various tropical and other extra-
European regions; he had ample materials placed at his disposal,
from the collections of Martius, Drége, Preiss, and other German
travellers, and from the herbaria of Hooker, Arnott, and Lindley
in this country, and he came to be regarded as the great autho-

18 MR. G@. BENTHAM ON GRAMINES.

rity for the determination of exotic Gramines. His ‘Agrosto-
graphia Brasiliensis’ is perhaps the best of all his works; and
his Graminee for the ‘Flora Africe australis’ is also very good.
His generic and subgeneric groups appear to me to be often
better, or at least more natural, than those of Kunth or Trinius ;
although they show in some degree that tendency to multiply
genera as well as species, which he afterwards carried to so great
an extent in Cyperacee, Laurinex, and Acanthaces. Moreover,
he worked up the grasses of each country separately, without
paying sufficient attention to the cosmopolitan nature of so many
species, which thus appear under different names in his different
works. Brown’s Australian Panicum semialatum, for instance,
is raised by Nees to the rank of a genus under the name of Cori-
dochloa in India, and that of Blufia in South Africa, without any
attempt at a comparison of the three plants.

The last general Enumeration of Gramines was that of Steudel,
who published in 1855 the first volume of his ‘Synopsis Plan-
tarum Glumacearum,’ the worst production of its kind I have
ever met with. He was an excellent mechanical compiler; his
‘Nomenclator Botanicus’ was a most useful work ; and if in the
Grasses he had confined himself to collecting all the published
species with references to or copies of their author’s characters or
descriptions, he would have rendered good service to the students
of the Order; but beyond that, as he was no botanist, he was
thoroughly incompetent for the task he had undertaken, When-
ever he met with a grass which he could not readily make out, he
set it down as new, with anew name, and a character so carelessly
drawn up as to render its identification hopeless without recourse
to the specimens themselves. Several of his new genera are well-
known species repeated in the ‘Synopsis’ under their published
names without recognition. A few, indeed, may have to be re-
tained ; but others, again, are founded upon the grossest errors,
as, for instance, where he describes as a caryopsis the larva
which had eaten up the ovary and taken its place in the enlarged
pericarp. Having, moreover, no idea of methodical arrangement,
his work is a perfect chaos.

Much has been done, however, for the elucidation of the Order
in local Floras. Already at the close of last century and the
commencement of the present one, several continental botanists
proposed new genera for anomalous European grasses; but these
were published in works which entered but little into general cir-

MR. G. BENTHAM ON GRAMINESX, 16

culation, and were overlooked by Beauvois, Persoon, Willdenow,
and other general systematists. Several of the same genera
have since been reestablished, but under other names which have
now been so long and so universally adopted, that they must be
considered as having acquired a right of prescription to overrule
the strict laws of priority. It would indeed be mere pedantry,
highly inconvenient to botanists, and so far detrimental to science,
now to substitute Blumenbachia for Sorghum, Fibichia for Cyno-
don, Santia for Polypogon, or Singlingia for Triodia. Since the
days of Kunth, Trinius, and Nees, the most important local re-
visions of Graminex are: Andersson’s ‘ Graminex Scandinavia,’
Parlatore’s first volume of his ‘Flora Italiana,’ Cosson and
Durieu’s Glumaceous volume of the great unfinished ‘Flore
d’Algérie,’ Doell’s Graminee for the great Brazilian Flora
founded by Martius, and Fournier’s Gramines for the Mexican
Flora he has undertaken ; besides more partial revisions by
Grisebach in his ‘Spicilegium Flore Rumelice et Bithynice,’ in the
fourth volume of Ledebour’s ‘ Flora Rossica,’ and in various con-
tributions to the Floras of extratropical South America, the West
Indies, the Himalayas, &c., and by Emile Desvaux in Claude
Gay’s ‘Chilian Flora,’ supplemented by new genera and species
published by Philippi in various papers on Chilian plants.
Andersson was a most acute observer, and had studied well the
northern grasses of the old world; but from want of access to
a sufficiently extensive library, his synonyms, especially when
treating of extra-Scandinavian species, are often very inaccurate.
Parlatore’s detailed monograph of Italian grasses is thoroughly
to be relied upon when the result of his own observations; but
unfortunately neither he nor Andersson sufficiently distinguished
the characters they had taken from other works from those they
had themselves verified. Old errors, for instance, in the de-
scriptions of the style or of the ripe fruit, which it is often very
difficult to ascertain from dried specimens, have been in several
instances repeated by both authors, sometimes in identical terms.
Both of them also, especially Andersson, show a great tendency to
the multiplication of genera and species. Cosson and Durieu's
‘Monograph of Algerian Grasses,’ comprising the chief portion
of those of the rich West-Mediterranean Flora, is a most valuable
treatise, both for methodical arrangement and specific distinc-
tions. Grisebach has also done much for the elucidation of oriental
Graminew. In Doell’s work I have been disappointed. In many

20 MB. G. BENTHAM ON GRAMINER.

instances I cannot approve of his distinctions or combinations of
genera or species. That may, however, be a matter of opinion
only; but in regard to several of the exceptional characters he
gives, such as the five lodicules of Pariana or the three of Aris-
tida, they have not been verified on reexamination ; in his spe-
cific names he has not unfrequently departed from the established
rules of nomenclature without giving any special reasons for so
doing; and there is a general carelessness in redaction showing,
for instance, on several occasions that when he had found reason
to modify his first ideas as to the limits of species, he had neglected
to revise his manuscript accordingly. He also makes frequent
use of the expression “partis nomine,’ the meaning of which
neither Munro nor myself, nor any of our classical friends to whom
we have applied, can make out. Eugéne Fournier’s ‘Enumeration
of Mexican Graminex’ is not yet published; but being already
printed off, and M. Fournier having obligingly supplied me with
a copy, I feel bound, in so far as I am concerned, to treat it as
having already taken date. He has had at his disposal rich col-
lections of the grasses of a country where they are perhaps more
local and varied even than in South Africa; and he has made good
use of these materials, although there is still much to be learnt
with regard to Mexican forms. We have at Kew several, not
only species but genera, which are not included in his work; and
there are not a few of his which I cannot recognize in our gene-
rally rich Kew collections. A further comparison is also required
with extra-Mexican genera and species, and especially with those
of extratropical South America. His genus Lesourdia, for in-
stance, had already been published for a southern species by
Philippi under the name of Scleropogon. His Trichloris is re-
presented in the south by two species separately recognized by
Munro and by Jean Gay as constituting a distinct genus, but
under names hitherto unpublished, which must therefore give way
to Fournier’s. Ina systematic point of view also his work would
have been much more useful if he had more frequently given the
characters of the tribes, genera, or other groups which he has
modified, instead of limiting himself to dichotomous keys. These
dichotomous keys, when carefully drawn up, are of the greatest
use as guides or indexes to direct the botanist where to look for
his plant, but are wholly insufficient for its identification either
generic or specific. For above sixty years I have had great
experience both in using and in making them. It was with the

MR. G. BENTITAM ON GRAMINESR. 21

aid of the admirable “Analyses”? in DeCandolle’s ‘Flore Fran-
gaise’ that I was enabled in 1817 and 1818 to learn botany without
any extraneous teaching; their principle was developed in the
‘Essay on Nomenclature and Classification’ which I published in
1823 as a French edition of Jeremy Bentham’s ‘ Chrestomathia,’
and I have introduced them more or less into all my local floras.
Tam thus well aware of the great difficulties in the way of draw-
ing them up satisfactorily, requiring much testing before their
final revision. They are chiefly useful where all, or nearly all,
the plants of a country or of a group are well known; and even
then they frequently require the repetition of the same plant
under different branches of the key. The best genera and other
groups are usually distinguished by a combination of characters,
to each one of which there may be occasional exceptions, and
these cannot be provided for in any key that presupposes limits
definitely marked out by single characters. Ag a result, there
are some of Fournier’s groups which are evidently good, but to
which we have no clue but that supplied by the species he includes
in each. The two genera or subgenera, for instance, into which
he divides Bouteloua, Lag. (Hutriana, Trin.), are natural and well
limited; but the only character he gives, the prolongation of the
rhachis of the spike beyond the last spikelet in the one and not
in the other, is in fact variable in both groups. Of others, again,
I can form no idea of the limits he proposes to assign them. In
Uniola, for instance, he admits species (unknown to me) which do
not appear from his description to have what we have been accus-
tomed to consider as an essential character of the genus, the four
to six empty glumes at the base of the spikelet. Where, there-
fore, I feel obliged to differ from him in the genus to which I
would refer a species, it may as often be from the inability to
ascertain what are his views as to the limits of a genus, as from
that difference of opinion which so frequently prevails amongst the
best of botanists.

In recent days, however, we had all been led to look up to my
much lamented friend the late General Munro as the one who
was to unravel the intricate web into which the order had become
involved. His ‘Monograph of Bambusew’ and various detached
papers and communications were instalments of great promise ;
he was known to have a thorough acquaintance with species, and
to have already formed a well-digested framework for genera and
tribes, an important sample of which he had given in the second

22 MR. G. BENTHAM ON GRAMINES.

edition of Harvey’s ‘ Genera of South-African Plants.’ He had
also amassed an immense number of notes on synonyms he had
verified, on points of structure he had ascertained, &c., as mate-
rials for the general work he was preparing for DeCandolle’s
Monographs. His death has extinguished all such hopes as we
had entertained; and although his notes, mostly dispersed in his
herbarium or in the gramineous books of his library, are now left
at our disposal at Kew, yet he had unfortunately not committed
to paper his ideas on the limits and distinctive characters of tribes,
genera, and subgenera not included in the South-African Vlora ;
and these I could only gather from his conversation and corre-
spondence. My own preparation for the work I have now under-
taken was chiefly the study of European grasses for my ‘ Hand-
book of the British Flora,’ and of Old- World Graminex generally
for the ‘ Flora Hongkongensis’ and ‘ Flora Australiensis,’ when 1
was in constant correspondence relating to them with General
Munro. Having now had to work also upon American forms and
to examine with more detail the South-European, Oriental, and
African ones, I have had to modify in some respects the views I
had expressed as to the relative importance and constancy of some
of the characters, and partially to rearrange some of the tribes
and subtribes, although the general principles of classification
which had been suggested by General Munro have only been con-
firmed by further experience.

Thave already, in my paper on the classification of Monocoty-
ledons (Journ. Linn. Soe. (Bot.) xv. p. 5138), entered so fully into my
reasons for adopting as to Graminex a terminology in accordance
with the observations of Mohl and in harmony with that followed
as to Cyperacee, that I need not repeat them on the present occa-
sion. I would only add a few words in further reply to the objec-
tion repeatedly made to me that the falling off together of the
flowering glume and palea (commonly called the two palew) en-
closing the fruit, isa strong evidence of their being really homo-
logous. But this is a mistake. A careful observation will show
that they never do both together fall away from the rhachilla or
axis of the spikelet ; it is the rhachilla itself that breaks up, a por-
tion of which always remains attached to the glume and palea and
keeps them together round the fruit. In most Panicaceer, espe-
cially in Andropogonee, the whole spikelet with the empty glumes
as well as the flowering one falls off with the fruit. In the majo-
rity of Poacez the disarticulation takes place between each twa

MR. G. BENTHAM ON GRAMINER. 23

flowering glumes, leaving the intervening portion either attached
to the glume next above it, when it is usually described as a callus
proceeding from the glume, or to the glume next below it, when
it is often half concealed between the keels of the palea and taken
no notice of; or if it be a continuation of the rhachilla above the
last glume, it is often termed a neuter or abortive flower. The
cases where the flowering glume really detaches itself ultimately
from the inarticulate and persistens rhachilla are very few, chiefly
in several species of Hragrostis, where the glume and caryopsis
fall away, leaving the palea and floral axis persistent on the rha-
chilla. In some cases the apparently terminal fruiting glume
enclosing the palea and caryopsis falls away without any percep-
tible portion of the rhachilla above or below it; but that arises
from the disarticulation taking place so close under it that the
fragment carried off is only that minute portion actually em-
braced by the base of the glume.

The homology of the glumes of Graminez, whether empty
or flowering, with those of Cyperacee may now be considered
as generally admitted; and a total absence of perianth in the
former order might not be regarded as improbable when we have
traced in Cyperacee its gradual reduction from the regular hexa-
merous perianth of Oreobolus to its absolute deficiency in Cyperus
and others. But we have in Graminex a new element on the
floral axis below the stamens and pistil or actual flower, in the
palea and lodicules, for which we cannot at once find any parallel
in other orders, and which have been very variously accounted
for. They have very recently been the subject of a very able
paper in Engler’s ‘ Botanische Jahrbiicher’ (i. p. 336) by Pro-
fessor Hackel of Vienna. He comes to the conclusion that the
palea and the pair of lodicules (when two only) are each of them
single, more or less bifid, organs, and that they and the third lodicule,
when present, must be regarded as two or three bracteoles inserted
alternately fore and aft on the floral axis below the flower. And
he has made out a good case in favour of his view, but perhaps
not an unanswerable one. The first objection that strikes one is
that the difficulty of finding any homologues in other orders is
by no means diminished. In other orders where bracteoles do
exist below the flower, they are usually lateral with reference to
the main axis, not fore and aft, never more than two, unless when
representing a continuation, as it were, of the sepals, and never
developed, to my knowledge, when the perianth is suppressed ;

24 MR. G. BENTHAM ON GRAMINER.

the bracts performing the functions of the deficient perianth
are always, I believe, on the main axis, like the glumes of Gra-
minee. Then, again, the perfect union of the two lobes of the
palea or of the two lodicules, or even the occasional development
of a single central nerve or central lobe, is no absolute proof that
they are not in fact double organs ; for where the segments of a
perianth are united in a tube or cup, the lateral nerves of two
adjoining segments (sepals or petals) often coalesce into a single
one which may protrude at the top into an intermediate tooth or
lobe. Hackel has well shown that the unity or duplicity is the
same in the case of the palea and of the two lodicules; but it 18
only conjecturally that he continues the parallel through the third
lodicule, which, when present, never shows any tendency to divi-
sion, and whose insertion is not perceptibly higher up than that
of the two others. It is quite true that it is often much smaller
than the other two, sometimes very minute; but in several spe-
cies of Stipa, in the majority of Bambusee &c., I have seen the
three quite equal and perfectly similar. The only instances I know
of more than three lodicules are those of Ochlandra, where they
are exceedingly irregular, and of Reynaudia, where I find four in
two pairs, as described and figured by Kunth; but then the outer
pair, although closely contiguous (on the opposite side of the
floral axis) to the upper ones, appear to me to represent the palea
which is otherwise deficient. The minute bodies above the lodi-
cules in the female flowers of Pariana, which Doell mistook for
additional lodicules, appear to me to be rudimentary staminodia ;
they are very minute and irregular, and not always to be found.

I have observed that the search for homologues to the palea and
lodicules in the Orders nearly allied to Graminee has met with
but little success. The only representation of the palea that I
can find is that mentioned in my above-quoted paper (Journ. Linn.
Soc. (Bot.) xv. p. 516), where it is compared with the hypogynous
scales of Hypoelytrum pungens and Platylepis ; and I find that in
some species of Hriocaulon (Flora Australiensis, vii. p. 190) the
perianth is composed of two outer segments inserted near the base
of the floral axis and two or three inner ones close under the
andreecium, or these inner ones occasionally deficient, the arrange-
ment passing gradually through other species to the normal two
contiguous series of two or three each. It might therefore be
suggested that the palea and lodicules of Graminew represent
perianth-segments of an outer and inner series, although I by no

MR. G. BENTHAM ON GRAMINED. 25

means pretend to assert it asa proved fact. If the suggestion be
confirmed, we might be justified in designating as a neutral flower
that in which the palea alone, or the palea and lodicules without
stamens or pistil, are developed; but we must not include in the
flower the bract or glume which subtends it.

In all cases the palea, whatever its origin, is called upon in con-
junction with the subtending glume to perform more or less of the
functions of the deficient or absent perianth, and thus acquires a
certain fixity of character, and requires mention in all full generic
characters. The lodicules, on the other hand, are generally rudi-
mentary representatives of suppressed organs having lost’all func-
tional powers, and their slight variations in form or consistency
are generally not even of specific importance, and they only re-
quire mention in generic characters in the few cases where they
have retained a greater and more constant development.

There is much of interest in the question cf the geographical
distribution of Grasses as compared with that of Orchidew, and in
the consideration of the causes which have produced the differ-
ences observed in the two Orders, amongst which perhaps the very
different agencies through which cross-fertilization is effected may
be most influential; these questions may have also more or less
bearing on tribual and generic arrangement; and there are nume-
rous observations which I should have been desirous of recording.
This, however, would lead to speculations which it would not be
safe to indulge in without a far more detailed and closer study of
ascertained facts than I have time to carry out; and I feel obliged
to confine myself ou the present occasion to the purely systematic
consideration of real or supposed affinities and diversities.

The division of the Order into tribes and subtribes is a matter
of exceptional difficulty. Whatever tribes have been proposed,
whatever characters have been assigned to them, there have
always been more or less ambiguous forms uniting them and
preventing the restricting them within absolutely definite limits.
We are obliged in Graminee, more perhaps than in any other
Order, to rely upon combinations of characters, allowing for
occasional exceptions in every one of our groups, preferring
those which experience has shown to present the fewest aberra-
tions. Following up these views, none of the general divisions
of the Order hitherto proposed have proved to be more natural
or more definite than Brown’s original primary one into two
great groups or suborders—Panicacee, in which the tendency to

26 ME. G. BENTHAM ON GRAMINEE.

imperfection is in the lower flowers of the spikelet ; and Poaceae,
in which the tendency is in the opposite direction. This indica-
tion of the principle kept in view is too indefinite to serve as a
practical character; but, combining it with that proposed by Munro
of the articulation in the axis of the spikelet being below the
spikelet itself (in the pedicel) in Panicacew, and above the lowest
glume or none in Poacex, the exceptional forms are reduced
to the lowest possible figure. This primary division, although
tacitly approved of by many partial agrostologists, has not been
generally adopted in systematic works, and many attempts have
been made to divide the Order according to more positive cha-
racters, but as yet with but little success.

Kunth entirely gave up Brown’s primary groups and divided
the Order into thirteen tribes, many of which were natural,
fairly defined by a combination of characters, and have been very
generally adopted. Others have been objected to on various
grounds. He attached too much importance to such characters
as the separation of the sexes or the increase in the number of
stamens, which are exceptional in different groups rather than
tribual distinctions ; in the general arrangement, his removal of
the Andropogonez to a distance from the Panicex is disapproved
of; and his describing flowers as actually existing when only theo-
retically imagined is sometimes misleading. Nees generally
adopted Kunth’s tribes, but improved the circumscription of
some of them, and added two or three small ones. Trinius never
completed his revised arrangement of the Order. Since the time,
however, of these great agrostologists, systems have been sketched
out which require a few words of notice.

Fries, followed by Andersson, proposed for a primary division
of Graminee that into Clisanthee, with the flower (i.e. the
flowering glume and palea) closed and the elongated styles pro-
truding at the apex, and Luryanthee, with the glume and palea
open at the time of flowering and the short styles protruding
laterally, This division is, however, practically useless, except
perhaps for the limited number of species that can be observed
in a living state. The flowers of most species open only for a
very short time, and in dried specimens are almost always closed.
The styles, again, are in many cases so exceedingly slender and
fugacious as to be very difficult to observe in dried specimens,
except in the bud, when they have not yet attained their full
development, or after fertilization, when they are withering away,

MR. G. BENTHAM ON GRAMINEZ, 27

The long styles, moreover, would place the majority of the sub-
tribe Seslerier, for instance, among Panicacex, when all their
other characters are those of Poacee; and the species are very
numerous in which, from the intermediate length of the styles,
or from both the lower smooth part and the stigmatic portion,
or the lower part alone, being described as styles, they are
differently characterized as long or short by different writers.

Fournier rejects both Brown’s and Fries’s primary divisions,
but proposes a new one founded on the position of the lowest
glume of the spikelet, next to the main axis in Chloridem and
Hordeacer, and averted from it or external in other tribes. But,
in the first place, this relative position cannot well be ascertained
in loosely paniculate Graminew, where there is so frequently a
slight, almost imperceptible torsion of the pedicel, and, in the
next place, in one-flowered spikelets it is often uncertain which
is to be regarded as the lowest glume. The total number of
glumes in the tribe Panicew, for instance, is variable, according
to the genus or section, two, three, or four; the lowest in
Reimaria, and in a few species of Paspalum, corresponds to the
second in the majority of Paspala and a few allied genera, and to
the thirdin Panicum. All these genera are included by Fournier,
as by all others, in one and the same tribe; and if so, are we to
regard as the outer glume the small outer one of Panicum, called
by some an extra bract, and an imaginary one in Paspalum and
its allies, or the outer one of Paspalum, which is second in Pani-
cum? Again, in one and the same genus, the relative position of
the outer glume and the main axis is not always constant, as, for
instance, in Paspalum, in Nees’s section Diyitarie (Emprosthion,
Doell, Anastrophus, Schlecht.), the outer glume and the flowering
one above it are external, whilst in the majority of the genus
they are turned towards the central rib of the main axis, and
yet the two groups are not distinguished by Fournier even as
sections.

Another character much insisted on of late years for tribual
distinction is still more uncertain, the adherence of the ripe grain
or caryopsis to the palea, as in Festuca, Bromus, &c. This is
usually very “conspicuous in the dry state, although even then
the grain is often only closely embraced by the palea, and when
moistened the adherence very generally disappears. The union
of the two is perhaps never truly organic, and in hot water I
have always found them readily separable without any tearing.

28 MR. G. BENTHAM ON GRAMINEX.

The consequence is that there are a considerable number of
species in which the grain has been described by some as ad-
herent and by others as free, and which have consequently been
transferred from one genus to another. Yet, if not taken too
absolutely, the character is sometimes a useful one, assisting, for
instance, in the arrangement of the genera of some of the sub-
tribes of the difficult tribe Festucex.

Considerable importance was attached by the earlier agrosto-
logists to the presence or absence of the awn on the back or
apex of the flowering glume; but this has subsequently been
found to be subject to great variations. The spiral twist, how-
ever, in the lower part of the awn in some genera is more con-
stant, and in the ‘Flora Australiensis’ I had taken it as an essential
character of some tribes or subtribes; but there are more ex-
ceptions than I was then aware of. The awn, when present,
is generally twisted in Andropogonee, Tristeginer, Agrostidee,
and Avenacee, and not in Panicew, Chloridee, Festucee, or
Hordee ; but it is sometimes very slightly so ina few species of
the latter group, and in the former tribes, where the awn is much
reduced, if there be any twist it is scarcely perceptible. In all
the tribes, also, the awn is occasionally, and in the straight-
awned ones frequently, altogether deficient ; and in some genera,
as in Stipa for instance, where it is usually twisted, there are
exceptional species in which it is straight or curved only. The
character must therefore generally be used with more or less
of reservation.

The partial or absolute separation of the sexes or the increase
in the number of stamens observed in a few genera have been
occasionally introduced amongst tribual characters; but further
observation has shown that they occur amongst Graminee of
very different affinities, and have thus proved to be often of no
more than generic value, although in one tribe, the Maydee, the
absolute unisexuality of the spikelets may be constant.

Differences in the size of the embryo, in the form of the so-
called scutellum on the caryopsis (indicative, apparently, of the
hilum of the seed), or in the longitudinal groove or cavity fre-
quently observable on the caryopsis, have been sometimes brought
forward as absolute generic, if not tribual, characters, and they
may often be really important; but we know, as yet, too little
about them to test their value fairly. Herbarium specimens

rarely supply ripe fruits, and they have been carefully observed

MR. G. BENTHAM ON GRAMINES. 29

and accurately described in comparatively few species. ‘The cha-
racters thus ascertained in a single one have been supposed to
belong necessarily to the whole genus; and when differences have
afterwards been found in some other species, it has at once been
generically separated, without ascertaining whether these dif-
ferences might not be reconciled or connected through other
species. Before, therefore, we can ascertain the real generic
value of characters which cannot be tested in herbarium speci-
mens, it is necessary that we should have them well and authen-
tically described in a much greater number of species from
actual observation. I have on several occasions had reason to
believe that, in long-detailed descriptions drawn up by accurate
botanists from dried specimens, the seminal characters have been
rather guessed at on theoretical grounds, than actually verified on
really ripe seeds.

Following out the views of General Munro as to the general
arrangement of the Order in as far as I have been able to ascer-
tain them, we have divided it into tribes and subtribes, of which
the following are the most prominent characters, omitting for
the present exceptional forms, which occur in almost all of
them :—

A. Panicacrs. Spicule cum pedicello infra glumas articulate,
flore fertili unico terminali, addito interdum inferiore masculo
v. sterili.

Tribus i. Paniceze. Spicule hermaphrodite, rarius abortu unisexuales,
spicate v. paniculate, rhachi inflorescentie inarticulata. Gluma
florens exaristata, fructifera indurata v. saltem exterioribus rigidior.

Tribus ii. Maydez. Spicule unisexuales, mascule terminales spicata
v. paniculata v. (in Pariana) foemineam circumdantes, foeminez in-
feriores spicatze, cum rhacheos internodio (excepta Zea) articulatim
secedentes.

Tribus iii. Oryzeze. Spicule hermaphrodite v. rarius unisexuales, pani-
culatz v. spicatie, rhachi inflorescentiz inarticulata. Gluma sub flore
summa. (palea?) uninervis v. carinata.

Tribus iv. Tristegineze. Spicule hermaphrodite, secus panicule ra-
mulos inarticulatos solitariz v. rarius gemine v. fasciculate, cum
pedicello articulate. Glume vacuz aristate v. muticz, florens hyalina
y. tenuiter membranacea, arista geniculata terminata v. mutica.

Tribus v. Zoysiew. Spiculea hermaphrodite v. nonnulle imperfecti,
cum rhachi inarticulata spice simplicis sigillatim v. fasciculatim ar-
ticulate. Gluma florens membranacea, sepius vacuis minor hyali-
naque,

LINN. JOURN.—-BOTANY, VOL. XIX. E

30 MR. G. BENTHAM ON GRAMINE.

Subtribus 1. Anthephorex. Spicule in pedicello 3-%, in fasciculum
deciduum conferte. Gluma florens nunc vacuis sublongior, nunc
brevior hyalina. ; ;

Subtribus 2. Euzoysieex. Spiculein pedicello solitarie, rarius gemine.
Gluma florens vacuis brevior, hyalina.

Tribus vi. Andropogonez. Spicule secus spice rhachin v. paniculee
ramulos, sepissime geminz v. terminales ternz, in quoque pari homo-
game v. heterogame. Gluma florens vacuis minor, hyalina, sepe
aristata,

.

B. Poaczem. Pedicellus infra glumas continuus. Rhachilla
supra glumas inferiores persistentes sepe articulata, ultra flores
fertiles producta, stipitiformis v. glumas vacuas v. flores imper-
fectos ferens, v. interdum flos fertilis more Panicacearum unicus

terminalis, sed cum gluma sua a vacuis persistentibus articulatim
secedens.

Tribus vii. Phalaridez. Flos hermaphroditus unicus, terminalis.
Glume 6 (v. 5 et palea) uninerves v. carinate.

Tribus viii. Agrostez. Spicule 1-florz, rhachilla ultra florem nuda v.
in setam v. stipitem producta.

Subtribus 1. Stipes. Panicula luaa v. irregulariter spiciformis. Gluma
florens arista sepius terminata, fructifera caryopsin arcte involvens.
Rhachilla ultra florem non producta.

Subtribus 2. Phleoidese. Panicula spiciformis densa, cylindracea v. sub-
globosa. Gluma florens mutica v. aristis 1-3 terminata fructifera
caryopsin laxe includens. Rhachilla interdum producta.

Subtribus 3. Sporobolex. Panicula laxa v. ad racemum reducta, raris-
sime spiciformis. Gluma florens mutica. Caryopsis demum sepius
glumis apertis subdenudata. Rhachiila non producta.

Subtribus 4. Euagrostex. Panicula varia, sepius lava. Gluma florens
sepius arista dorsali instructa, rarissime mutica.
laze inclusa. Rhachilla sepe producta.

Tribus ix. Isachneze. Spicule zqualiter biflore.
tice. Rhachilla ultra flores non producta.

Tribus x. Aveneze. Spicule bi- v. pluriflore, sepius paniculate. Glume
florentes arista dorsali v. interdum terminali sepissime instructa.
Rhachilla ultra tlores sepius producta.

Tribus xi. Chlorideze. Spicule uni- y. pluriflore, secus rhachin spica-
rum unilateralium biseriatim sessiles, secund:.

Tribus xii. Festuceze. Spicule bi- y. plurifloree, varie paniculate v.
rarius racemose. Glume florentes mutice y. aristis terminate,

Subtribus 1. Pappophoree. Glume florentes plurinerves tri- pluri-
aristate, v. absque aristis quadriloba.

Subtribus 2. Triodiex. Glume florentes uni- v. trinerves, tridentate, tri-
fide v. triaristate,

Caryopsis gluma

Glumee seepius mu-

MR. G. BENTHAM ON GRAMINER. 31

Subtribus 3. Arundines. Rhachilla sub glumis florentibus longe pilosa.
Subtribus 4. Sesleriez. Inflorescentia spiciformis v. capituliformis, basi
glumis vacuis v. spicis sterilibus sepius stipata, Stylus v. rami sepius
longi tenues.
Subtribus 5. Eragrostex. Glume florentes trinerves. Cetera normalia.
Subtribus 6. Melices. Glume florentes tri- v. plurinerves, superiores
due v. plures vacue, semet involventes.
Subtribus 7. Centothecex. Folia plana, lanceolata v. ovata, inter venas
transverse venulosa. Glume florentes quinque- v. plurinerves.
Subtribus 8. Eufestucerx. Glume florentes quinque- v. plurinerves.
Cetera normalia,
Tribus xii. Hordeeze. Spicule uni: v. pluriflore, ad dentes seu excava-
tiones rhacheos spice simplicis sessiles.
Subtribus 1. Tritices. Spicule ad nodos solitarie, tri- v. plurifiore,
rarius biflore.
Subtribus 2. Leptureze. Spicule ad nodos solitarie, uni- v. biflore. Spica
tenuis.
Subtribus 3. Elymez. Spicule ad nodos gemine v. plures collaterales.
Tribus xiv. Bambuse . Gramina elata, sepius basi saltem lignosa.
Folia plana, seepissime cum vagina articulata. Spicule uni- v. pluri-
flore. Lodicule szpius 3. Stamina 3, 4, v. plura.
Subtribus 1. Arundinariese. Stamina 3. Palea bicarinata. Pericarpium
tenue, semint adnatum.
Subtribus 2. Eubambuseze. StaminaG. Paleabicarinata. Pericarpium
tenue, semini adnatum.
Subtribus 3. Dendrocalamez. Stamina 6. Palea bicarinata, Peri-
carpium crustaceum v. carnosum, a semine liberum.
Subtribus 4. Melocanneze. Stamina 6 v. plura. Palea 0 nisi glumis
simillima. Pericarpium crustaceum v. carnosum, a semine liberum.

T now proceed to a more detailed revision of the several tribes,
subtribes, and genera, in the order in which I have worked them
up for the forthcoming part of our ‘ Genera Plantarum,’ to which
I must refer for the technical characters and references, as well
as for the synoptical clavis of the genera.

Series A. PANICACEZ.

This first main division of Graminex is very fairly defined by
the combination of two characters—the articulation of the pedicel
below the spikelet or cluster of spikelets, and the single fertile
flower apparently terminal, with or without a single male or
sterile one below it. Where either of these two characters fails,
the plant should be referred to Poaceee.

The articulation of the pedicel is usually immediately below

E2

32 MR. G. BENTHAM ON GRAMINEZ.

the lowest glume, leaving, as the spikelet falls away, a slight
dilatation or callosity at the apex of the persistent portion.
Sometimes it is not easily observed at the time of flowering, but
becomes more marked as the fruit ripens. A similar marked
articulation has not hitherto been observed in Poacez, except in
Fingerhuthia. There are also a very few cases where the lowest
glumes are reduced to slight callosities, or are so rudimentary as
to render it difficult to say whether the articulation is in the
pedicel or in the rhachilla. In the Cenchrus group of the tribe
Panicer, in the subtribe Anthephorer of Zoysiew, and in some
Andropogonew, the articulation is not under each spikelet, but
under a little cluster of two or more spikelets ; and in Maydee it
is the rhachis of the spike which disarticulates under each female
spikelet. In Graminee generally, however, the articulation,
whether of the rhachis, of the pedicel, or of the rhachilla, is
usually under the fertile spikelets or flowers only; under the
males it is apt to be very obscure or quite obsolete.

The fertile flower is above spoken of as only apparently ter-
minal, because the presence of the palea and a slight obliquity
tend to show that the floral axis is not really the continuation of
the rhachilla, but, as in Poaceee, a secondary or axillary branch.
Doell says, indeed, that a continuation of the rhachilla behind
the palea hay been observed in a species of Panicum; but I have
never succeeded in meeting with it in any Panicacew. In the
tribe Oryzew, where there is no two-nerved palea, it may still re-
main a matter of doubt whether the floral axis is or is not distinct
from the rhachilla—whether the uppermost scale is a glume on
the rhachilla or a palea at the base of the floral axis. The pre-
sence or absence of a central nerve is not an absolute test ; for it
is occasionally, though very rarely, absent in the lower glumes.

Panicacee have never more than four glumes, the uppermost
one usually enclosing or subtending the fertile flower, though in
some Andropogonex it is excessively reduced or even quite
obsolete or rudimentary. The next under it may be empty like
the lower ones, or may enclose a palea, a rudimentary flower, or
a perfect male flower, and in Beckmannia, and a very few species
or individuals of Setaria and Panicum, this lower flower may be
hermaphrodite, but usually, if not always, sterile. The two lower
glumes when present are always empty. Where the spikelets
are unisexual, the females have only the single terminal flower,
the males most frequently two flowers, both with perfect
stamens.

MR. G, BENTHAM ON GRAMINER. 33

The tribes composing the series of Panicacee run much into
each other, and have been very variously extended or reduced.
We have adopted the following six, as having appeared to us to
be rather better defined than the smaller or larger ones that have
been proposed.

Tribe I. Panton.

The principal character of the Panices, considered as a tribe
of Panicacee, consists in the hardening of the fruiting glume.
In several of the smaller genera, however, and even in some
species of Panicum itself, it is membranous, but usually larger
than the outer ones, and forming the chief covering of the fruit,
never hyaline or much reduced as in Andropogoner. Oryzopsis,
Milium, and their allies, which were formerly included in Panicex,
have been transferred to Agrostidez on account of the persistent
lower glumes below the articulation. Among the other general
characters of the tribe, the inarticulate rhachis of inflorescence is
constant except in Stenotaphrum, where, however, the articulation
is very tardy and not constant, so that it has often been denied.
The flowering glume never bears the twisted awn, so general in
Andropogonee and Tristeginee, although in Hriochloa and a very
few species of Panicum its obtuse apex has a short, erect, almost
dorsal point; the awns of Oplismenus, Chetaria, the section
Echinochloa of Panicum, &c. are straight and terminate one or
more of the empty glumes only. The fertile flower terminating
the spikelet is, in the normal genera, either perfectly hermaphro-
dite, or, at any rate, as far as I have observed, has staminodia
round the pistil. It is only in a few of the abnormal genera
added to the tribe that there are strictly female spikelets.

The normal genera of the tribe may be distributed in four
rather distinct groups, though scarcely marked enough to be
raised to the rank of subtribes ; and to these we would add a few
more or less abnormal genera, but little connected with each
other, but all apparently more nearly allied to Panices than to
any other tribe.

In the first group, or Panicee proper, we have distinguished
eleven genera—a number somewhat arbitrary ; for much might be
said in favour either of uniting the whole into one vast genus
Panicum, or of dividing them still further, as some have proposed,
into about twice as many as those here adopted, the distinctive
characters being often either very uncertain, or such as are not
universally recognized as generic in the Order.

34 MR, G. BENTHAM ON GRAMINEE.

1. Rermanta, Fligge—This old-established and universally
acknowledged genus has generally been limited to two tropical
and subtropical American species, with a peculiar slender habit
and inflorescence, and characterized by having only one empty
glume below the flowering one, and by the constant reduction of
the number of stamens to two. It has since, however, been as-
certained that several species which cannot well be separated from
Paspalum have only a single lower empty glume ; and Doell has
distinguished Reimaria chiefly by the reduction of the stamens,
together with the form of the spikelets more acuminate and
more closely appressed to the rhachis than in any Paspalum. He
has added, under the name of R. aberrans, a third species, which,
with a more vigorous habit, rather invalidates the natural di-
stinction from Paspalum, but has all the characters of Reimaria ;
and Munro recognizes a fourth species, allied to #. aberrans, but
with only two, or at most three, spikes to the panicle and a
much thicker rhachis, in the Florida plant distributed by Curtis
with the number 3566 as Paspalum vaginatum, but probably not
the one entered under that name in Chapman’s ‘ Flora of the
Southern United States.’ It occurs also in Wright’s Cuban
collection under un. 3854, and may be characterized as R. oli-
gostachya, Munro, spicis in pedunculo 2 rarius 8 (nec 6-15),
rhachi dilatata spiculis sublatiore. The true Paspalum vagina-
tum, Sw., is a synonym of P. distichum, Linn.

2. Paspanum, Linn., ranks among the large genera of tropical
Gramineze, and in respect of the greater number of species is a
natural one, readily distinguished from Panicwm by the inflo-
rescence and by the technical character of the deficiency of the
small lowest glume. It is now, however, ascertained that neither
character is quite constant. A few Panica of the section Bra-
chiaria have the inflorescence of Paspalum ; and the lowest glume
is frequently reduced to a small callus, or is entirely deficient in
the section Digitaria ; and the consequence has been, that several
species have been referred by some botanists to the one genus
and by others to the other. These ambiguous species appear,
however, to be best placed in Paniewm ; and all true Paspala have
the spikelets sessile or nearly so, in two or four rows along the
lower or outer side of the rhachis of the spikes or simple branches
of the panicle, and they show no trace of the small lowest glume
of Panicum. ‘Thus defined, the number of species may be esti-
mated at about 160, by far the greater proportion of them tro-

MR. G. BENTHAM ON GRAMINE. 35

pical American, a few of which are also generally spread over
the warmer regions of the Old World, especially P. distichwm, Linn.
(P. vaginatum, Sw.), which reaches southern Europe as an intro-
duced weed. Scarcely five species can be regarded as belonging
exclusively to the Old World. The above estimate of the total
number is founded chiefly on the investigations of Munro, who
had nearly completed the working-up of the genus, and has left
full descriptions with diagnoses and synonymy of 188 species,
besides a few that he had left for further inquiry. Steudel
enumerates 262 species, but nearly half of them have proved to
be mere synonyms or very slight varieties. Doell describes in
detail 105 Brazilian species ; but some of them are what I cannot
consider as really distinct; and his own views of them were any
thing but stable, as there are several which he at one time re-
ferred to one species and later transferred to another, forgetting
to eliminate them from their former place, thus :—

Gardner, n. 2354, is repeated under P. malacophyllum and
P. subsesquiglwme.

Hostmann, n. 658, under P. densiflorum and P. cespitosum.

P. distachyum, Salzmann, u. 667, under P. pumilum and P.
divergens.

Gardner, n. 8496 and 3497, under P. maculosum and P. notatum.

Gardner, n. 2975, under P. vaginatum, P. tropicwm, and P.
Jilifolium.

P. cespitosum, Hochst., n. 1548,

P. amazonicun, Trin., and

P. humile, Steud.,

Digitaria uniflora, Salzm., n. 659, | under P. platycaulon and

under P. plicatulum and P.
dissectum.

and Spruce, n. 679, P. furcatum.
P. surinamense, Hochst., n. 1283, under P. furcatwm and P.
scoparium.

Spruce, n. 30, under P. chrysodactylon and P. chrysoblephare.

Fournier enumerates 40 Mexican species, of which thirteen
are described as new ; but he is, in Graminew, generally disposed
to admit as distinct species forms which I perfectly agree with
Munro in regarding as slight varieties, corresponding to what so
many local European botanists describe as critical species.

With regard to the subdivision of the genus, Trinius, in his
several revisions, distributed the species chiefly according to the
size of the spikelets, which, however much it may affect the
general aspect of the species, is in many cases far too uncertain
a character to be practically useful. Nees, in his ‘ Agrostologia
Brasiliensis, proposed six sections, which Doell reduced to four,

36 MR. G. BENTHAM ON GRAMINEE.

Munro, though he had so nearly completed his descriptions of
species, and often indicated the sections to which he referred
them, had not yet definitively grouped them, leaving his manu-
scripts, for convenience of reference, in alphabetical order. We
have adopted three sections, founded on Nees’s, which appear to us
well defined by positive characters—EHupaspalum, Cabrera, and
Anastrophus, subdividing the first, and largest, into four groups
or subsections, Anachyris, Opisthion, Pseudoceresia, and Ceresia,
much less marked in their outlines, but generally speaking fairly
natural.

Eupaspalum comprises the great majority of the species, and is
distinguished by the spikelets strictly secund along the rhachis of
the spikes, with the back of the flowering glume and of the lower
empty one (when present) turned outwards—that is, away from
the rhachis or from its midrib; whilst in Anastrophus, which in-
cludes the remainder of the genus except the monotypic Cabrera,
the spikelets are almost distichous, and the back of the flowering
glume and of the lower empty one turned towards the midrib of
the rhachis. This distinction was specially relied upon by Nees
under the terms spicule adverse and spicule inverse, and followed
up by Doell. I¢ is not alluded to by Fournier with regard to the
Mexican Paspala ; but, if I understand correctly his words (Gram.
Mex. p. vii), it nearly corresponds to the character he proposes
for the primary division of Gramines.

Anachyris, the first subsection of Hupaspalum, is a purely arti-
ficial one, characterized solely by the having only a single empty
glume below the flowering one. It was first proposed as a genus
by Nees for the Paspalum malacophyllum, Trin., which has all the
habit and floral and other characters of Paspalwm except this
single one; and Fournier, apparently on this account, transfers
it to the tribe Oryzee. Doell, however, reduces it to a section
of Paspalum under the name of Eyremachyrion, associating with it
a few other species, some of them evidently more nearly allied to
corresponding species of the section Opisthion than to each other.
And even the technical character is not always constant; for in
P. (Eremachyrion) sesquiglume, Doell, a species closely allied to
P. (Opisthion) maritimum, Trin., I frequently find a minute outer
glume; and, again, P. pallidum and P. candidum, H. B. K., both
of which Doell places in Lremachyrion, are scarcely to be distin-
guished from each other except by the lowest empty glume absent
in the one, present in the other, as originally pointed out by

MR. G. BENTHAM ON GRAMINER. 37

Kunth. Nees describes the palea (upper palea) of the typical
Anachyris paspalodes or Paspalum malacophyllum as 8-nerved ;
Fournier says it is 1-nerved. The species is very variable as to
the size of the spikelets, the hairs or sete on the rhachis of the
spike, &c.; but in all the specimens I have examined I have
uniformly found the palea normally 2-nerved.

Opisthion, proposed by Doell as a section of Paspalum, is our
second subsection of Eupaspalum. It includes all the typical
Paspala with two lower empty glumes, and the rhachis of the
spikes not dilated. The species are numerous and varied, but
scarcely reducible to distinct groups.

Pseudoceresia is a subsectional name I should propose for the
genus Ceresia as understood by Elliott and other North-American
botanists. In it the rhachis of the spikes is more or less dilated
and concave, but green and herbaceous throughout, and the spike-
lets are small and glabrous or nearly so. The species are few,
including P. stoloniferum, Bosc, P. repens, Berg.,and their allies.
Ceresia is the name we would reserve for our fourth subsection,
being the genus Ceresia as originally established by Persoon, in
which the dilated rhachis of the spikes is bordered by a coloured
or smooth membranous margin, and the half-enclosed spikelets
are larger than in Pseudoceresia and densely ciliate. Besides
several Brazilian and other tropical species, it includes the Mexi-
can P. cymbiforme, Fourn.

Cabrera, our second section of Paspalum, is limited to the single
P. aurewm, H. B. K. (not of Trinius), forming Lagasca’s genus
Cabrera, in which the direction of the spikelets is nearly that of
Anastrophus ; but instead of being marginal on each side of the
rhachis, they are deeply embedded in alternate cavities on each
side of the midrib, on the outer or lower side of that rhachis. This
remarkable arrangement is very well described by Lagasca, who
was a most accurate botanist. His ‘Nova Genera et Species
Plantarum,’ forming part of the ‘Elenchus Horti Matritensis,’ is
a model for the clearness and conciseness of the characters given,
which are most thoroughly to be depended upon. The work is
quoted by Nees and by Doell, but evidently at second hand; had
they really read it, and had they studied Kunth’s good figure and
description, they could never have given to the P. awreum the
new name of P. immersum, or have transferred the synonym of
Cabrera chrysoblepharis, Lag., to the P. exasperatum, Nees, or to
the supposed distinct P. chrysoblepharis, Doell, both of them at

38 MR. G. BENTHAM ON GRAMINEE.

complete variance with Lagasca’s description. The genus Axo-
nopus, Beauy., sometimes given as a synonym of Cabrera, because
Beauvois had suggested that P. awreum might possibly be a con-
gener, was founded on various heterogeneous species of Paspalum
and Panicum; and the name has to be wholly expunged.

Anastrophus, our third section, was proposed as a genus under

that name by Schlechtendahl, and includes Nees’s section Digi-
tariee or Doell’s Emprosthion. It is characterized by the posi-
tion of the spikelets on the alternate margins of the narrow,
somewhat flexuose rhachis of the spike, so as to be rather di-
stichous than secund, and by their direction, the back of the flower-
ing glume and of the lower empty one being turned outwards or
away from the rhachis. The spikes are also generally several
close together at the end of the peduncle, as in the section Digi-
taria of Panicum, suggesting to Nees his sectional name, which,
however, is inconvenient as being adjective in form, and too liable
to be confounded with the true Digitaria. Some of the species
have, like Cabrera, long cilia on the spikes, but have otherwise all
the characters of Anastrophus, of which they might form a sub-
section under the name of Lappagopsis, given by Steudel to the
P. dissitiflorum, Trin., which he proposed as a distinct genus. The
several species which we would include in the subsection show a
curious diversity in the position of the cilia: in P. fastigiatum,
Nees, they are long on the empty glumes, none on the rhachis;
in P. senescens, Nees, short on the empty glumes, long on the
rhachis; in P. dissitiflorum, Trin., long both on the rhachis and
on the empty glumes; and in a few other species, referred by
Nees and by Doell to Cabrera, although without the peculiar
characters of Lagasca’s genus, the rhachis alone is fringed with
long cilia, the glumes having none.

Paspalum saccharoides, Trin., referred by Kunth to Panicum,
is one of those small-flowered species which seem to connect Pas-
palum with Panicum (Digitaria), whilst the long silky hairs of the
spikes and the consistence of the glumes show an approach to the
Andropogonee (Saccharee). The arrangement of the spikelets
along the rhachis, the number of glumes, &c. show a nearer
affinity to Paspalum than to any other genus.

3. Anrumnantia, Beauv. (Awlavanthus, Ell.), was founded upon
two North-American species, with the hairy inflorescence and
membranous glumes of the section Trichachne of Panicum, but
without the small lowest glume of that genus; and the second

MR. G. BENTHAM ON GRAMINES. 39

glume (corresponding to the third of Panicwm) usually encloses
a palea ora male flower—a circumstance unusual in the Order,
where the exposed glumesare almost always empty. From these
I cannot separate generically the South-American Leptocoryphium,
Nees, which, besides some slight specific characters, only differs
from the North-American species in the second glume being con-
stantly, instead of occasionally only, empty. The genus Anthe-
nantia thus constituted includes three species—A. villosa, Beauv.
(Aulaxanthus ciliatus, Ell., Panicum ignoratum, Kunth), A. rufa,
Benth. (Aulaxanthus rufus, Ell., Panicum rufum, Kunth), and
A. lanata, Benth. (Paspalum lanatum, H. B. K., Milium lanatum,
Kunth, Leptocoryphium lanatum and L. molle, Nees).

4, AmpHicarpum, Kunth, with spikelets unisexual by abortion
and a peculiar inflorescence, remains limited to the single North-
American species on which the genus was founded.

5. Errocutoa, H. B. K. (a name having the right of priority
over Gidipachne, Link, and Helopus, Trin.), has the habit rather of
the section Brachiaria of Panicum than of Paspalum, but wants the
small lower glume of the former genus, and differs generally from
both in a peculiar callous thickening of the pedicel at the articu-
tion. There are, however, a very few species with more or less
of this callosity, which on other accounts cannot well be separated
from Panicum. The flowering glume has also the peculiar point
on the obtuse apex observable in Panicum helopus, Trin., and in a
few others, and supposed to characterize a section or genus Uro-
chloa. It is, however, an uncertain character, both in Eriochloa
and in Panicum. Nearly twenty supposed species of Hriochloa
have been described ; but the greater number of them are scarcely
even varieties of the H. polystachya, H. B. K., which is widely
spread over the warmer regions of the Old as well as the New
World, and known under the various names of Z. punctata, E.
annulata, &c. There appear also to be at least four really distinct
species—Z. distachya, H. B. K., and 2. grandiflora (Helopus,
Trin.) from tropical America, £. trichopus, Hochst., from tropical
Africa, and Z. villosa, Kunth, from eastern Asia.

6. Bucxmanyta, Host, is a single species, ranging from eastern
Europe across Russian Asia to North America. It has been
usually placed in Phalaridex, a tribe with which it appears to me
to have but little connection. The habit and inflorescence are
those of Panicum colonum; but it is exceptional in Panicew as
having both the flowers hermaphrodite; the lowest flower is,

40 MR. G. BENTHAM ON GRAMINEZ.

however, as far as I have observed, usually sterile ; and a similar

character is to be found in some species or varieties of Setaria,

and very rarely in Panicum itself, next to which the genus appears.
to be best placed. The synonym Joachimia, Ten., given by Kunth,
was a name intended for it by Tenore, but I believe never actually

published. Tenore figures the plant as a Beckmannia.

7. Panicum, Linn., after deducting Ichnanthus, Oplismenus,
Setaria, and several smaller genera, remains one of the larger, and
probably the largest, among tropical Grasses, and is still in many
respects polymorphous. In habit and inflorescence it may be
confounded sometimes with Paspalum, sometimes with Arundi-
nella, or even with some Agrostew. Generally speaking, it may
be easily recognized by technical characters ; but the most marked,
the very small size of the lowest empty glume, is not quite con-
stant; for in a few species this glume is wholly deficient as in
Paspalum, whilst in a few others it is of the size of the second
glume ; the hardening also of the fruiting glume and palea is in
some species very slight. There is nothing, however, sufficiently
definite or constant in these exceptional species to mark them out
as intermediate genera; and here, as in so many other cases of
large genera of Cyperacee and Gramines, we must admit the
existence of forms which must be placed in one or the other of
allied genera from considerations of convenience rather than of
strict character. Taking the genus Paniewm within the limits we
have ascribed to it, nearly 800 supposed species have been pub-
lished: Steudel enumerates 716; Doell has 184 Brazilian ones,
Fournier 97 Mexican, Nees 44 South-African; I described 54
Australian ones ; and they are rather numerous in tropical Africa
and Asia; but a considerable number are repeated in several or
even in all of these Floras, and a large proportion of Steudel’s
species are mere synonyms or blunders. The total number of
fairly distinct species can therefore scarcely be estimated at much
above 250. These have been variously grouped, chiefly according
to their inflorescence ; and no less than eighteen supposed genera
have been at different times separated from it, but are now re-
united, either as being founded on insufficient, uncertain, or even
mistaken characters, or as being, in our opinion, more conveniently
regarded as sections than as genera. But, even as sections, their
limits are often as far from being absolutely definite as are those
of the whole genus. The following eleven are those which have
appeared to be the most distinct; but they are all more or legs

MR. G. BENTHAM ON GRAMINE®. 41

connected by intermediate forms, and several of them would pro-
bably be modified, and may hereafter be much improved, by a
closer study of species than I have at present been able to bestow
upon them.

(1) Digitaria. Spikelets usually small and in alternate pairs or
clusters along one side of the simple spike-like branches of the
panicle; those of each pair or cluster unequally pedicellate, or
one of them almost sessile, and the lowest glume often very
minute or sometimes quite deficient. This section was proposed
ag a distinct genus in Walter’s ‘ Flora Carolinensis’ under the
name of Syntherisma, and by Richard, in Persoon’s ‘Synopsis,’
under that of Digitaria, and is still maintained as such by many
botanists. It was founded originally on the cosmopolitan weed
Panicum sanguinale, Linn., in which the spike-like branches of
the panicle are clustered at the end of the peduncle like those of
Cynodon and some other Chloride. There are now, however,
nearly forty species to be included in the group, in many of which
the spikes or branches are distant along the peduncle, as in
Schedonnardus, Gymnopogon, Leptochloa, &c., among Chlorides.
From this tribe the structure of the pedicellate spikelets and their
articulation always keep them perfectly distinct ; but there is a
series of small-flowered species, Including the Australian and
Asiatic P. parviflorum, Br., P. tenuiflorum, Br. (Paspalum brevi-
folium, Fligge), and Paspalum minutiflorum, Steud., and two or
three from South Africa, which have been almost equally well
placed by some in Paspalum, by others in Panicum. As in some
species allied to P. sanguinale, and even in some varieties of
P. sanguinale itself, the minute outer glume is frequently abso-
lutely deficient. The more pedicellate spikelets and the occa-
sional, however rare, appearance of the outer glume may justify
the placing these species rather in Panicum than in Paspalum, to
which I referred them in the ‘Flora Australiensis.’ P. platy-
carphum, Trin., from Bonin Island, with all the characters of true
Digitaria, is remarkable for the dilated membranous rhachis of
the spike-like branches as in the section Ceresia of Paspalum.

(2) Trichachne. In this section, distinguished as a genus under
that name by Nees and others, the branches of the panicle are
simple as in Digitaria, but usually few, loose, scattered along the
pedunele, and erect. The glumes are all, or the second ones
alone, ciliate or clothed with soft hairs as in the section Pricho-
lena; and the fruiting glume is not much hardened. The species

42, MR. G. BENTHAM ON GRAMINES.

are few: P. semialatum, Br., is widely spread over the Old World,
for I am unable to distinguish the Asiatic Coridochloa, Nees, and
the South-African Blujfia, Nees, from Brown’s Australian species ;
P. Gayanum, Kunth, is confined to tropical Africa; P. leuco-
pheum, H. B. K., is frequent under various names in the tropical
and subtropical regions of the New and the Old World. Itisa
very variable species; and specimens gathered at different stages
of development look very different from each cther, but are not
separable into marked varieties. It was included by Beauvois in
his genus Urochloa, and appears to-have been the type of the
proposed genera Acicarpha, Raddi, Hriachne, Philippi, and Holo-
setum and Mesosetum, Steud., and is probably the principal ele-
ment of Presl’s supposed genus Alloteropsis.

(3) Diplaria. This section is proposed for a few American species
with a simple terminal spike-like inflorescence. The spikelets
are sessile along the rhachis in two rows and distichous, as in the
section Anastrophus of Paspalum, from which Diplaria differs
technically in the presence of the small outer glume characteristic
of Panicum. It comprises P. rottboellioides, H. B. K., P. exaratum
and P. ferrugineum, Trin., P. pappophorum, Nees, and a few others.

(4) Thrasya, distinguished as a genus by Kunth, has a simple
terminal spike-like inflorescence asin Diplaria ; but the rhachis is
more or less dilated as in the section Ceresia of Paspalum, and
the spikelets, sessile along the midrib, although really alternate
and biseriate, have all the appearance of being in a single row.
The species are few, all American, and include, besides the ori-
ginal Thrasya paspaloides, Kunth, the P. ansatum, Trin., which
is scarcely specifically distinct from it, P. thrasyoides, Trin., P.
petreum, Trin., and perhaps two or three others. The P. petreum
forms the genus Zylothrasya of Doell, which he characterizes by
a callous thickening of the pedicel like that of Zriochloa; but the
plant is in all other respects too closely allied to the typical
Thrasya to be generically separated, and the callosities are slightly
prominent in various species of Panicum.

(5) Harpostachys. The inflorescence is again simple and spike-
like; but the spike is more or less falcate, with the spikelets
crowded in two or four rows along one side of the slender rhachis,
as in Chloride, and the common peduncles are usually long and
often clustered two or three together in the upper axils. To
this section belong P. monostachyum, H. B. K., P. decumbens, R.
et Schult., and P. subfalcatum, Doell.

MR. G. BENTHAM ON GRAMINES. 43

The genus Dimorphostachys of Fournier is founded upon the
above P. monostachyum and some other American species, which we
should refer to the sections Digitaria or Brachiaria, but which
he connects generically by the small lowest glume being more
developed or differently shaped in one spikelet of each pair than
in the others; but the difference is often exceedingly slight,
and the character so little connected with any other or with
habit, that it seems difficult to attach any more than specific
importance to it.

(6) Brachiaria. This section, sometimes referred to as Paspaloid
Panica, comprises a large number of species both from the New
and the Old World, in which the inflorescence is that which is
regarded as specially characteristic of Paspalum: the panicle con-
sists of a number of spike-like simple branches, distributed along
a simple common peduncle ; but the small lowest glume of Pani-
cum is always present. If we regard only such typical species as
P. flavidum or P. fiuitans of Retz, or P. paspaloides of Persoon,
the section appears a most distinct one; but, on the other hand,
several such species as P. adspersum, Trin., P. argenteum, Br.,
P. Petiveri, Trin., P. polyphyllum, Br., &c. so closely connect it
with some of the sparingly-flowered species of Hupanicum, as to
make it impossible to draw a precise line of demarcation between
the two. Amongst these intermediate forms, Paractenum, pro-
posed as a genus by Beauvois, appears to be only a starved state
of P. gracile, Br.

P. helopus, Trin., bears on the obtuse apex of the flowering
glume a short point, like that of most’species of Eriochioa, and
was therefore joined by Beauvois to the P. (ZLrichachne) semi-
alatum, Br., to form his genus Urochloa; but the two are in other
respects too dissimilar to be united in one section, and P. helopus
appears to be altogether a true Brachiaria.

(7) Echinochloa, was regarded by Beauvois as a distinct genus,
founded chiefly on two very widely-spread and most variable
species, P. colonum, Linn., and P. crus-galli, Linn., the former
often cultivated, the latter a most abundant tropical and sub-
tropical weed. Both have nearly the inflorescence of the section
Brachiaria but they are coarser plants, with the spikelets densely
crowded ou the partial spikes or branches of the panicle, and the
second and third empty glumes, in the one rarely, in the other
very generally, terminating in long awns. It was probably on
this account that Kunth united Beauvois’s Kehinochloa with his

44 MR. G@. BENTHAM ON GRAMINES.

Oplismenus ; but the development of the awn has now been shown
to be so frequently uncertain in one and the same species of
Graminez, that the character has quite lost the absolute import-
ance once attributed to it by Beauvois and others, and Echinochloa
is generally admitted only as a slightly distinct section of Panicum.
The true Oplismenus may, however, be well maintained as a sepa-
rate genus, to which I shall presently refer.

(8) Ptychophyllum has been well worked up as a very distinct
section of Panicum by A. Braun. It comprises P. plicatum, Lam.,
from the Old World, P. swleatwm, Aubl., from America, and a
few others, which, with a peculiar foliage, have more or less of
sete in the panicle, which seem to connect them with Setaria.
On examination, however, these sete will be found in Ptychophyl-
lum to be merely the setiform tips of the ultimate spikelet-bearing
branches of the panicle, whilst the bristles or sete of Setaria are
abortive branchlets, forming a kind of involucre below the spike-
lets. The remaining floral characters of Ptychophyllwm are entirely
those of the loosely-panicled species of Hupanicum.

(9) Hymenachne of Beauvois, often retained wholly or partially
as a genus, comprises a small number of species both from the
New and the Old World, in which the small, very numerous
spikelets are usually crowded in a long narrow cylindrical spike-
like panicle. In the typical species, P. mywrus, Linn., the spike-
lets are rather acuminate and the fruiting glume scarcely hardens;
in P. indicum, Linn., and others the spikelets are small, and
quite those of a large number of true Panica.

(10) Eupanicum. After deducting the nine preceding sections
and the succeeding Tricholena, which have all some distinguishing
peculiarity, there remain a large number of species strictly normal
in the structure of their awnless spikelets, and connected together
by their more or less spreading panicle, the spikelets, on short
or ou slender pedicels, clustered or scattered along its simple or
divided branches. These species, in number not far from two
hundred, may vary much in the size of the spikelets, in the degree
of development of the panicle, and in other minor points, but
seem little capable of being classed in distinct subsections. They
form Trinius’s two sections Virgaria and Miliaria, characterized
by the branches of the panicle being angular in the one, terete in
the other—a distinction which I have been quite unable to follow
out, at least in the dried specimens. All I have been able to
suggest has been their distribution into seven groups or series,

MR. G. BENTHAM ON GRAMINES, 45

vaguely distinguished chiefly by their inflorescence and general
habit. Amongst the somewhat exceptional species are P. unci-
natum, Trin. (Echinolena polystachya, H. B. K.), in which the
three empty glumes are nearly equal to each other, though shorter
than the flowering ones, and P. pterygodium, Trin. (forming the
genus Otachyriwm, Nees), in which the two lower empty glumes
are about equal, but shorter than the third. Inall the others the
lowest empty glume is much the shortest. Coleatenia, from
extratropical South America, is proposed asa genus by Grisebach
as having dicecious flowers. I have not seen any specimen; but
from his description it seems to be in all other respects a true
Panicwm (Eupanicum) ; and as he has only seen the male, evidently
with the flowers still young, he may have overlooked the pistil, or
its abortion may not be constant. At any rate that character
standing alone can scarcely be sufficient to separate it generically.
Several of the cultivated Millets are species of Hupanicum with
large, loose, often nodding panicles.

(11) Lricholena (including Nees’s genus Rhynchelytrum), raised
by Parlatore and some others to the rank of a genus, has the loose
panicle of HLupanicum; but the fruiting glumes are not much
hardened, and the whole inflorescence is ciliate with long hairs
as in Trichachne, on which account the oldest known species, the
widely-spread P. Teneriffe, was originally published as a Saccha~
rum. There are now about fifteen species known, chiefly South-
African ; but one, the above-mentioned P. Teneriffe, extends to
the Mediterranean region, two are East-Indian, and two or three
South-American. Rhynchelytrum, Steud., is a different genus
from Nees’s, and belongs to the Tristeginee.

8. Icuyantuvs, Beauy., is so closely allied in habit and general
character to some species of Panicum (Eupanicum) that it is
perhaps rather in deference to the authority of all the principal
recent agrostologists, than from any conviction of our own, that
we retain itasa distinct genus. The character is a purely technical
one—a thin hyaline auricle or wing to the rhachilla on each side
close under the flowering glume, as is observed in some species
of Cyperus. In the species forming the section Macropteris of
Doell these auricles are often more than half as long as the glume
itself; in I. longiflora (Panicum longiflorum, Trin.) they are very
small, but prominent ; in Z. pallens and its allies, forming Doell’s
section Micropteris, they are often scarcely perceptible, and
Fournier has restored these species to Panicum, though Munro

LINN. JOURN.— BOTANY, VOL. XIX. F

46 MR. G. BENTHAM ON GRAMINES.

keeps them up as Ichnanthus. Two species, I. Hoffmanseggit,
Doell, and I. oplismenioides, Munro, are remarkable for the long
spreading hairs, which give them a very peculiar aspect. There
are altogether about twenty species, all tropical American.

9. OptisuEnus, Beauv. (Orthopogon, Br.), though very near the
section Brachiaria of Panicum, appears to bea natural genus, and
is well characterized by the greater development of the lowest
empty glume, which is, moreover, always awned, whilst in Panicum
it is much smaller than the others and always unawned. Kunth
adopted the genus, but, relying on the awns alone, united with it
Echinochloa, in which the proportions of the glumes are the ordi-
nary ones of Panicum, and which I have referred to above as a
section of Panicwm. Fournier adopts Kunth’s view. Steudel
and Doell both reduce the whole to Panicum. The true Opilis-
ment are widely spread over the warmer regions both of the New
and the Old World, and are variable as to the number and length
of the spikes or panicle-branches, &c. Some botanists adopt
above thirty species, others reduce the whole to varieties of a
single one; it is probable that some three or four may be fairly
distinguished as species. Hekaterosachne of Steudel is one of
the common forms of Oplismenus.

10. Cuarrum, Nees, to which Doell has properly referred

Berchtoldia of Pres] as a second species, has nearly the spikelets
of Oplismenus, to which Kunth reduces it, the outer glumes
being much more developed and awned than the flowering ones ;
but, besides some minor points, the inflorescence appears quite
different enough to justify the maintaining it as a distinct genus.
Doell considers it as a section of Panicum, with two species, one
Brazilian, the other Mexican. Fournier retains the genus Berch-
toldia for the Mexican one, without comparing it with Chetium,
and adds two supposed new Mexican species: the one, B. holci-
formis, judging from the specimens he quotes, is one of the large
coarse forms of Panicum (Echinochloa), very nearly allied to, if
not varieties of, P. erus-galli; the other, B. oplismenoides, is
unknown to me, but must from his description be referable also
to Echinochloa.

11. Srrarta, Beauv., was included by the older authors in
Panicum, and has been restored to that genus as a section by
Steudel and by Doell, but is retained by most modern botanists
as a well-marked natural genus, easily recognized by the dense
spike-like panicle usually bristling with numerous sete issuing

MR. G. BENTHAM ON GRAMINES. 47

from the pedicels below the spikelets. These sete are not
epidermal like the rigid hairs of many Graminee, but, as in
Pennisetum, are supposed to be abortive branchlets of the
panicle, differing, however, from those of the latter genus by
being inserted below the articulation of the pedicel, so as to
remain persistent after the fall of the spikelet. The species are
very variable, and a large number have been described as distinct ;
they appear, however, to be reducible to about ten, three of which
are common weeds over a great part of the civilized world, and a
fourth (8. ifalica) has been much cultivated as one of the Millets
of the Mediterranean region and the Levant. The genus was
first fully characterized by Beauvois in his ‘ Agrostographie,’
chiefly from the above-mentioned common weeds; but he had pre-
viously published and figured, in his Flora of Oware and Benin,
under the name of Setaria longiseta, a plant which, as far as I can
judge without seeing the specimen, proves to be no Setaria at all,
but the Pennisetum (Beckeropsis) unisetum, to which I shall pre-
sently refer. A few species or varieties of Setaria—one, for
instance, gathered by Hildebrandt in the Sandwich Islands, allied
to S. viridis, another, not uncommon in the Mexicano-Texan region,
allied to 8. italica—have, like the variety of S. glauca figured by
Trinius, t. 195, the lower flower hermaphrodite as well as the upper
one, which is quite exceptional throughout all genera of Panicew
except Beckmannia. Izxophorus, Schlecht. in Linnea, xxxi. 420,
was founded as a genus on Urochloa uniseta, Presl, a Mexican
grass which we do not identify in our collections; but Trinius
refers it to Panicum and Fournier to Setaria, with which Schlech-
tendal’s description agrees very fairly.

In our second or Cenchrus group of Panicee we include four
genera, chiefly tropical or subtropical, characterized by the
so-called involucre of bristles surrounding each spikelet or
sometimes each cluster of two or three spikelets ; this involucre,
supposed to represent abortive branchlets of the inflorescence,
being placed above the articulation of the pedicel, always falls
away with the spikelets; the spikelets themselves are quite those
of Panicum, the inflorescence usually a simple spike raceme or
spikelike panicle, rarely a loose panicle of two or more pedunculate
spikes. 12. Cencurvs itself, as reduced from the original Linnean
genus, consists of about a dozen species, both from the New and
the Old World, two or three of them of very wide geographical
range, all characterized by the numerous bristles of the involucres

F2

48 MR. G. BENTHAM ON GRAMINEA.

hardened and frequently more or less united at the base, the
inner ones often broad and scale-like. In some specimens, how-
ever, of C. calyculatus, Cav., and its allies the hardening appears
so slight as to bring the genus into very close connexion with
Pennisetum.

13. Pennisetum, Pers., the principal genus of the group, would
now contain about forty species, chiefly African, amongst which
two or three extend to the Mediterranean region, tropical or central
Asia, or tropical America, and a very few may be endemic in Asia,
Australia, or America. It has been at various times proposed to
separate several genera from it, and two or three of these have been
pretty generally adopted ; but they pass so gradually one into the
other, and their chief characters, derived from the hairiness or
numbers of the involucral bristles, are so little in accord with any
other characters or habit, that the several following groups can
scarcely be considered even as definite sections. Pennisetum
itself has been restricted to those species in which the bristles are
numerous and some or al] of them more or less hairy ; whilst those
in which the whole of the bristles are perfectly glabrous form the
genus Gymnotric, Beauv. But however easy this distinction may
appear at first sight, it is neither natural nor always definite. In
a few African species proposed by Figari and De Notaris as their
genus Eriochate, the whole of the setz are densely woolly-plumose;
in some of the commoner species numerous outer sete of each
involucre are glabrous, and as many or more or fewer of the inner
ones are hairy. In P. flaccidum, Munro, from East India, and
P. Benthamianum, Steud., from tropical Africa, amongst very
numerous glabrous ones there are generally only two or three
hairy ones, or sometimes none at all, thus forming a gradual con-
nexion with the true species of Gymnotrix, where the sete are
always quite glabrous; and there is nothing else whatever to
distinguish the two series even as marked sections. P. lanatum,
Klotzsch, is a remarkable Himalayan species, in which the
involucral bristles are few, sometimes reduced to a single long
rigid branched one, either plumose or glabrous, showing well the
true nature of the involucre of the genus. Penicillaria, Willd.,
often still retained as a genus, was founded upon a plant frequently
cultivated in the Indo-African regions, which may at first sight
appear to be abundantly distinct. The long dense cylindrical
spike or spike-like panicle is often above a foot long and an inch
in diameter, although in other cultivated specimens not above

MR. G. BENTHAM ON GRAMINER. 49

half that size. The involucres sometimes remain persistent after
the spikelets have fallen away, and the filiform styles are remark-
ably long; but many cultivated specimens and some East-African
ones, possibly wild, offer so much variety in these respects, some
passing quite into normal Penniseta, that it seems probable that
the peculiarities of habit have arisen from long cultivation. The
long styles united at the base occur in other species, amongst
which P. (Gymnotrix) macrostachyum, Brongn., has on that account
been proposed by Hasskarl as a genus, under the name of Sericura.
Amphocheta of Andersson is a Galapagos species of the Gymnotrix
group, with small spikelets in slender pedunculate spikes, forming
a loosely paniculate inflorescence, very different from that which
characterizes the greater number of Penniseta, but closely con-
nected with them through the several varieties of P. (Gymnotria)
tristachyum, Kunth. In P. (Gymnotrix) unisetum, Nees, an
African species proposed as a genus by Figari and De Notaris
under the name of Beckeropsis, this peculiar inflorescence is
carried still further, and the involucre is sometimes reduced to a
single bristle (always above the articulation and falling away with
the spikelet), though I usually find 2, 8, or even more bristles.
It is probable that the plant figured by Beauvois as Setaria lougi-
seta is this same species of Pennisetum. Steudel’s proposed genera
Catatherophora and Oxyanthe are normal species of Pennisetum
(Gymnotriz).

14, Praciosretum, Benth.,is a single Australian species, which
I characterized as a genus chiefly from its peculiar inflorescence
and habit, which prevented my retaining it in Pennisetum without
an extension of the generic character beyond what I felt justitied
in proposing.

15. Paratuertia, Griseb., is a single West-Indian species,
which proves to be identical with the Brazilian plant since pub-
lished by Doell as a section of Leptachyrium of Panicum, but
which is evidently more nearly related to Pennisetum. The
inflorescence is a simple spike-like panicle, of which the numerous
short articulate branchlets or pedicels are continued beyond the
single spikelets into long awns or bristles, which fall away with
the spikelet like the involucres of Pennisetum, thus forming in
some sort a connexion between the Cenchrus group of genera and
the following one.

Our third or Chameraphis group of Panicee consists of seven
small genera, loosely connected by a character which may be con-

50 MR. G. BENTHAM ON GRAMINED.

sidered as rather artificial than natural, but which I believe to be
constant. The spikelets are nearly those of Panicum, but with
the fruiting glume usually less hardened; the inflorescence is
nearly that of the paspaloid Panica or of the Chlorides, but dis-
tinguished from the former by the rhachis of the partial spikes or
fascicles or branches of the panicle being produced beyond the
spikelets into a more or less rigid point. From Chlorides the
articulation of the pedicel below the spikelet always separates the
present group. The genera are:—16. Ecurtnotmna, Desv., a
single tropical American species (H. scabra), which has quite the
rigid single spike of some Chlorides, but the spikelets of Panicez
intermixed with barren ones, on which account Rudge originally
figured the plant as a Cenchrus. The loosely paniculate species
added to Echinolena by Kunth have been rightly restored to
Panicum by Trinius. 17. Coammraruts, Br., four Australian or
tropical Asiatic species, fully described in my‘ Flora Australiensis.’
18. Spartina, Schreb. (Zrachynotia, Mich., Limnetis, Pers., Pon-
celetia, Thou., Solenachne, Steud.), five or six European, African,
or American species, chiefly maritime, has been usually placed’
amongst Chlorides ; but the spikelets themselves containing a
single terminal flower, and the articulation of their pedicels, are
quite those of Panicez, not of Chloridee. 19. XzRoocutoa, Br.,
three Australian species, 20. SrenotaPuRuM, Trin. (Diastem-
anthe, Steud.), two or three tropical maritime species, 21. PHyYL-
LORHACHIS, Trimen, a single one from Angola, and 22. THuaREA,
Pers. (Ornithocephalochloa, Kurz), also a single maritime species
from the shores of the Indian and South-Pacifie oceans, are all
perfectly isolated genera whose peculiarities have been well
pointed out. Stenotaphrwm is the only genus I know in the
tribe Panicez which has the rhachis of the inflorescence articu-
late ; but this can usually not be perceived except in an advanced
state, and has been denied by some botanists. I have already
alluded (Journ. Linn. Soc., Bot. xvii. 196) to Kunth’s mistake,
which induced him to alter Persoon’s name Thuarea (abridged
from Thouars’s then MS. name of Iicrothuarea) to Thouarea.
There remain seven very anomalous genera, but little connected
with each other, and still less with any other genera of Graminez,
but which have all more of the general character of Panicese than
of any other tribe. They have all been well defined and illus-
trated, and require no more than a bare enumeration on the
present occasion. They are :—23. Sprvirex, Linn., three Austra-

MR. G. BENTHAM ON GRAMINER. 51

lian species, of which one extends to New Zealand and New
Caledonia, with a fourth from the coasts of tropical Asia closely
allied to one of the Australian ones; 24. Onyra, Linn., about
twenty species, of which one is tropical African, the remainder
tropical American, including as a section Lithachne, Beauv.
(Strephium, Schrad., Raddia, Bertol.); 25. Puarus, Linn., five
American species; 26. Lupraspts, Br., three or four tropical species
from Africa, Asia, or Australia, a genus nearly allied to, but per-
fectly distinct from, Pharus ; 27. Lyqeum, Linn., a single maritime
species from the Mediterranean region; 28. StREPTOcHzTA,
Schrad. (Lepideilema, Trin.), and 29. ANOMOCHLOA, Brongn.,

both single Brazilian species.

Tribe II. Maypra.

The grasses composing this tribe are usually erect and tall, with
flat, long or broad leaves, the spikelets always unisexual, the
males, in all except Pariana, in the upper part of the plant or of
the inflorescences, the females at the base or in the lower axils,
the grain, in all except Zea, enclosed in a hard stony case, formed
variously of an outer glume or of a subtending bract. "Where
there are several fruiting spikelets in one inflorescence they are
superposed, and each one falls away separately with the internode
to which it is attached, the rhachis of the spike disarticulating at
each node. The male spikelets either wither away or remain
persistent above at the end of the stem or on the top of the
uppermost fruiting spikelet. The tribe is thus perfectly well
defined and quite distinct from any other ; and the eight following
genera of which it is composed, all tropical or American, and
mostly small or monotypic, are likewise marked by positive cha-
racters. ,

1. Partawa, Aubl., an American genus of about ten species, is
in mauy respects anomalous. The females, as in the other genera,
are single at each node of the articulate inflorescence; but the
male spikelets, instead of forming a terminal panicle, surround
the female at each node and fall away with it. The stamens
are also indefinite in number, ten to twenty in the spikelets
examined, but Nees found as many as forty; whilst in all the
other genera of the tribe there are only the normal three,
Doell describes the female flower as having five lodicules; but
here there is probably a mistake. I have never been able to
see more than three, which are rather large; but there are

52 MR. G@. BENTHAM ON GRAMINER.

sometimes within them two or three very minute scales, which
may possibly be rudimentary staminodia. Doell has also pro-
posed to separate generically, under the name of Eremites, a
Brazilian plant which, from the single spike I have seen as well
as from his description and figure, appears to be no more than
a starved state of some true Pariana.

2. Corx, Linn. (Lithagrostis, Gertn.), contains three or four
East-Indian species closely allied to each other, one of which,
the common “Job’s tears,” is widely spread over the warmer
regions both of the New and the Old World, but in many places
of comparatively modern introduction. The hard covering of the
fruit here consists of the sheath of a subtending bract, the
withered glumes as well as the internode of the rhachis remaining
entirely enclosed within it.

8. Potyroca, Br. (Cyathorhachis, Nees), three or four tropical
Asiatic species in which the stony case of the fruit is formed by
the outer empty glume, which is completely closed over the
remainder of the spikelet as well as the internode to which it is
attached. The species are :—(1) P. bracteata, Br. (Coix heteroclita,
Roxb.), spicis masculis terminalibus ramosis, inferioribus an-
drogynis v. foemineis plerisque simplicibus, glumis exaristatis:
(2) P. Wallichiana (Cyathorhachis Wallichiana, Nees), spicis
masculis terminalibus ramosis, inferioribus androgynis v. foemineis
plerisque simplicibus, spicularum mascularum gluma exteriore
longe tenuiterque aristata: (8) P. macrophylla, sp. u., spicis
longis (omnibus?) androgynis simplicibus, glumis acuminatis
exaristatis ; folia adsunt 2-pedalia, 2 poll. lata, spice 4-6-polli-
cares: from the Louisiade Archipelago (MacGillivray).

4, CuronacHNE, Br., contains three species from tropical Asia
or Australia, in which the hardened fruit-case is formed, as in
Polytoca, of the outer empty glume, but the internode of the
rhachis, instead of being completely enclosed within it, is em-
braced only by its thickened margins, and is seen lying as it
were in a groove of the fruit-case.

5. Screracune, Br., is a single Javan species, with the
fruit nearly of Chionachne, but with a different habit, and the
hardened outer glume is produced beyond the fruit into an open
membranous appendage.

6. Trirsacum, Linn., consists of two or three American species
with the terminal male inflorescence usually more branched than
in the preceding Asiatic genera, approaching that of Euchlena

MR. G. BENTHAM ON GRAMINE®, 53

and Zea; and the hardened fruit-case is formed partly only by
the outer glume, and partly also by the broad thickened and
hardened internode of the rhachis.

7. Evontawa, Schrad. (Reana, Brign.), has, like Zea, the ter-
minal male inflorescence paniculate with numerous spikelets, and
the female spikes in the lower axils wrapped up in broad bracts,
from which are protruded the long filiform styles; but, as in the
preceding genera, the female spikelets are within each bract
superposed in a single row on the articulate rhachis of the single
spike. The affinity to Zea appears to be recognized in the
country ; for specimens have been received from Schaffner pur-
porting to be known as “ wild maize.”

8. Zea, Linn. (Mays, Gertn.).—This most important, widely
diffused, and most striking grass is only known in a cultivated
state, or perhaps as an escape from cultivation. With most of the
general characters of the tribe to which it gives its name, it
is exceptional not only in that tribe, but in the whole Order,
by the manner in which its numerous female spikelets are densely
packed in several vertical rows round a central spongy or corky
axis. How far this arrangement may have gradually arisen after
so many centuries of cultivation can only be a matter of conjec-
ture. Its gradual progress cannot be traced through the nume-
rous cultivated varieties, many of them described as species in
Bonafous’s splendidly itlustrated monograph ; and the idea that
some of them are wild indigenous forms must be traced to the
insufficiency of the observations recorded by travellers.

Tribe III. Ornyzez.

This tribe, as originally constituted, was loosely characterized,
chiefly by uniflorous spikelets and stamens more than three—a
character more or less dispersed through various different tribes ;
and several of the genera included in it by Kunth have since been
rejected. The close affinity of Oryzez and Phalarex has also been
recognized, though the limits of the latter tribe also have been very
unsettled. In the ‘ Flora Australiensis’ J had united the two as
an intermediate tribe, connecting, as it were, the two great primary
series of Panicacee and Poacex; but upon thewhole it seems better
to separate them as tribes technically distinct, but representative
of each other in the two great series. The essential character of
both resides in having the scale immediately under the single ter-
minal perfect flower keeled or 1-nerved like the glumes, so as to

54 MR. G. BENTHAM ON GRAMINER.

leave it uncertain whether it is a glume or a palea—that is, whether
it be attached to the end of the rhachilla or primary axis of the
spikelet, or to a secondary or floral axis reduced to a mere point.
There are theoretical reasons in favour of both explanations, and
actual observation is insufficient for determining the point. The
first of these views has appeared to me the most plausible; and I
have accordingly in my diagnoses and descriptions treated the
scale in question as the flowering glume, and considered the
palea as deficient, as it certainly is in some Andropogonee and
Agrosteew. In this view the technical distinction between the
two tribes would be, that the Oryzew have 2, 4, or rarely 3 glumes,
all above the articulation of the pedicel, and the Phalarew 4, 6, or
rarely 5 glumes, the lowest pair persistent below the articulation
of the rhachilla. Oryzez thus characterized may be thought as a
whole to be a rather artificial tribe; but they are divisible into
two much more natural groups or subtribes—Zizaniee, tropical or
American genera, often semiaquatic plants, with a loose in-
florescence and stamens often, but not always, more than three ;
and Alopecuree, European or temperate Asiatic or African genera,
with a dense spike-like inflorescence and stamens never more than
three.

Zizaniew includes the following eight genera :—

1, Hyprocutoa, Beauv., a single species from Carolina, and
there apparently rare, differing from Zizania chiefly in the inflo-
rescence reduced to few-flowered spikes, of which the terminal
one male and pedunculate, the lower ones female and sessile in
the axils.

2. Zrizanta, Linn., comprises two species, or according to others
two genera, each with two or more species. Asa whole, the genus
is a natural one, well characterized by the unisexual spikelets in
an androgynous panicle, each one with only two glumes and the
males with six stamens. The typical Z aquatica, Linn. (Hydro-
pyrum, Link), has the lower part of the panicle more spreading
and male, and the upper part narrow and female; it is widely
spread over North America, and includes the East-Russian and
Japanese Z. latifolia, which is absolutely identical with some
North-American specimens. The other species, Z. miliacea,
Kunth (Zizaniopsis, Doell), has the male and female spikelets
more mixed in the panicle, the awns shorter, the styles more
connate, and the grain broader—characters which appear to me
quite insufficient for generic distinction. It is a North-American

MR, G. BENTHAM ON GRAMINER. §5-

plant, and possibly also South-American, if Sello’s single specimen
described by Trinius and figured by Doell as Z. microstacyhs,
Nees, was really from South Brazil. I see nothing in the figure
or description to distinguish it from Z. miliacea.

8. Luzrora, Juss., has, like Zizania, unisexual spikelets with
only two glumes; but the spikelets are smaller, not awned, the
styles short and quite distinct, and there are usually more than
six stamens in the males. Six species are known from tropical
America or the southern States of North America. The relative
arrangement of the males and females varies as in Zizania. In
the typical ZL. peruviana, Juss. (L. brasiliensis, Moric.), in L. ala-
bamensis, Chapm., and in an apparently unpublished Guiana
species, both sexes are in terminal panicles, but on distinct stems.
In L. Spruceana, Benth., described by Doell (figured by G. F. W.
Meyer as L. peruviana, but not Jussieu’s plant), the males are in
a, terminal panicle, whilst the females are in the lower axils of
the same stem, as they are also said to be in L. longivalvis, Doell,
a Brazilian plant which I have not seen. In the proposed genus
Caryochloa, Trin. (Arrozia, Schrad.), also Brazilian, the males and
females are in the same panicle, the former in the upper, the latter
in the lower part. The stamens in this species appear also to be
always six only, which only occasionally occurs in the others ; but
the other characters are entirely those of Luziola, to which I
should unite the Caryochloa as L. micrantha (Arrozia micrantha,
Schrad.).

4. Poramopuita, Br., if we include in it AMaltebrunia, Kunth,
is a natural genus of three species, connecting in some measure
Zizania, of which it has the habit, with Oryza, of which it has the
small setaceous or acuminate outer glumes. In the typical
P. parviflora, Br., from Australia, the spikelets are more or less
polygamous, though the greater number appear to be her-
maphrodite; in P. leersioides (Maltebrunia leersioides, Kunth)
from Madagascar, and in P. prehensilis (Maltebrunia prehensilis,
Nees) from South Africa, they are usually all, or nearly all, her-
maphrodite. Kunth also distinguishes Potonoptila from Malte-
brunia as having two flowers to the spikelet, a character not
mentioned by Brown and which I have been unable to verify.
The spikelets figured by Kunth, Rev. Gram. t. 5. figs. 1, 2, & 5,
must be very rare and probably abnormal; I have searched in
vain for them both in Brown’s and in Beckler’s specimens.

5. Hyeroruiza, Nees (Potamochloa, Griff.), is a single Hast-

56 MR. G@. BENTHAM ON GRAMINE SD.

Indian semiaquatic species nearly allied to Zizania, but quite
distinct in its hermaphrodite flowers and other characters.

6. Oryza, Linn. (Padia, Zoll. and Mor.), an Asiatic genus, of
which the typical species, the well-known Rice, appears to be
really indigenous in Australia as well as in Hast India; but it
has been so much cultivated from time immemorial, that it isnow
found apparently wild in various parts of Africa and America.
It has produced a large number of different forms, nearly twenty
of which have been published as substantive species, all of which,
or nearly all, are reduced by others to varieties of O. sativa. The
Himalayan O. coarctata, Griff., appears, however, to have more
positive characters; and possibly two or three others may be
maintained as fairly established species.

7. Leersia, Swartz (Homalocenchrus, Mieg., Ehrartia, Wigg.,
Asprella, Schreb., Blepharochloa, Endl.), is essentially American ;
but the two commonest species—L. hexandra in tropical, L. ory-
zoides in more temperate regions—are widely spread also over the
Old World, and had probably long been so before the civilized
communication between the two continents. The genus is closely
connected with the Asiatic Oryza; but, besides the apparent
diversity in geographical origin, the smaller spikelets with thinner
glumes and the general inflorescence give to Leersia a different
aspect, and, in technical character, the want of the two small outer
glumes may justify its retention as a distinct genus. It is true
that those who unite it with Oryza maintain that these outer
glumes are represented by a cartilaginous ring at the base of the
spikelet ; but this ring is often so slight as to be rather imaginary
than real, and never more than what is observable in Eriochloa
and some other Graminew, where no such theoretical explanation
is wanted or attempted.

8. Acuimya, Griseb., is a single Cuban species, which the
author compares with the Australian Microlena; but the want of
any glumes below the articulation places it in Oryzee, not in
Phalaridee. It is in some other respects allied to Oryza and
Leersia; but the peculiar inflorescence, the form and proportion
of the glumes, &c. readily distinguish it. Grisebach found only
a single stamen in the flower, a character which I have no means
of testing, the spikelets in our specimens haying already lost
their stamens.

The Alopecuroid group of Oryzez consists of four genera :—

9. Brcxera, Fresen., two or three Abyssinian species, in some

MR. G, BENTHAM ON GRAMINES. 57

respects intermediate between the two groups. The structure of
the spikelets, with the two outer glumes very minute or deficient,
connects them with the preceding genera; whilst the spicate
inflorescence and three stamens are nearer those of Alopecurus,
although the spikes are much more slender and several on the
same stem, on long slender peduncles. The genus is confined to
those of the §@ of Steudel’s ‘Synopsis ;’ the species arranged
under § 6 have a very different structure, and form the section
Beckeropsis of Pennisetum.

10. Crypsis, Ait. (Antitragus, Gertn.), must be limited to the
original C. aculeata, which alone has the characters of the tribe.
All the other species usually referred to it have the 2-nerved palea
and other characters of the Agrostew, and were well separated by
Host under the name of Heleochloa. It is true that some short-
spiked varieties of Heleochloa schanoides have very much of the
aspect of C. aculeata; but besides the structure of the spikelets
and the articulation of the rhachilla, they are readily distinguished
by the rhachis of the spike, which is linear and cylindrical, not flat
as in Crypsis. ;

11. Cornucorra, Linn., is a single Oriental species, very near
Crypsis, but well characterized by the peculiar inflorescence and
by the form of the fruiting spike and peduncle, which has supplied
the generic name.

12. Atopscurus, Linn., including Colobachne, Beauy., and
Tozzettia, Savi, is a well-known and perfectly definite European
and temperate Asiatic genus, with the habit nearly of Phleum and
the structure of the spikelets that of Oryzew. Above forty sup-
posed species have been enumerated ; but at least balf of them
must be regarded as trifling varieties of the two or three com-
monest species, which have now, and perhaps from remote times,
spread over a great part of the civilized world.

Tribe LV. TristeGine x.

This tribe, first proposed by Nees, has been fully adopted and
much extended by Munro, and now consists of thirteen genera,
which had been variously scattered in Panicew, Andropogones,
and Agrostex, and are really more or less connected with the three
tribes. They differ from Panicee and approach Andropogonee in
the thin, often hyaline texture of the fruiting glume and palea, and
by the frequent presence of a slender, often bent awn on the
flowering glume. From Andropogonex they are chiefly separated

58 MR. G, BENTHAM ON GRAMINES.

by their inflorescence; the spikelets are singly scattered or
clustered along the inarticulate branches of the panicle or, in the
very few cases where they are in pairs, the two of each pair are
perfectly: similar. Tristeginee are distinguished from Agrostes
by the characters which separate the two primary series Panicacess
and Poacew: ‘The tribual name was given by Nees from the genus
Melinis, which he published as Tristegis, believing it to be new ;
and although its identity with Beauvois’s Melinis has since been
established, it does not seem worth while now to alter the tribual
name, which has been pretty generally adopted. Of the thirteen
following genera, several of them common to the New aud the
Old World, the first four, with three glumes to each spikelet, are
temperate or subtropical, the following nine all tropical, with four
glumes to the spikelet.

1. TuuxBeRrtia is anew name I have been compelled to substitute
for Greenia of Nuttall or Sclerachne of Torrey, both of which had
been preoccupied. The genus is limited to two North-American
species which Steudel has proposed to unite with Limnas; but
they differ essentially from it in the awn of the flowering glume
terminal, not dorsal, in the distinct styles, and other characters
besides habit. I have named the genus after G. Thurber, who has
much studied North-American Graminee and worked them up
for 8. Watson’s Californian Flora. The genus formerly dedi-
cated to him by Asa Gray has since proved not to be distinct from
Gossypium, to which it has been reunited by the author himself.

2. Limwas is a single perfectly distinct species from East-
Russian Asia, well described and figured by Trinius.

3. Potyeogon, Desf., a genus readily known by its dense in-
florescence and the long awns of its empty glumes, is one of those
which interferes in some measure with general classification. It
has usually been placed in Agrostes; but the very decided arti-
culation of the pedicel removes it from that tribe to the Triste-
ginew, where in many respects it is allied to Garnotia. It consists
of about ten species, dispersed over the temperate regions both of
the northern and the southern hemisphere, one of them almost
cosmopolitan, but they are rare within the tropics. It was first
published by Savi under the name of Santia in the Memoirs of the
Italian Society of Science, a publication which had so little circula-
tion that the name has not found its way into standard works, and
that of Desfontaines has now been so long and so generally in use
in all countries, that it would only create useless confusion now to

MR. G. BENTHAM ON GRAMINER. 59

take up Savi’s. The Mexican P. elongatus, H. B. K., which is
Presl’s genus Wowodworskya (first described and figured by him
under the name of Raspailia), has the pedicels, although clavate
as in the rest of the genus, yet less decidedly articulate, thus
forming some real connexion with the Agrostes.

4, Garnotia, Brongn. (Miquelia, Nees, Berghausia, Endl.),
which sometimes comes near to some forms of Polypogon, has, on
the other hand, the spikelets in pairs on the inarticulate branches
of the panicle as in Miscanthus, and thus very closely connects
Tristeginew with Andropogones. It has, however, none of the
long hairs on the rhachilla so common in Andropogonesx, and
cannot well be removed far from Arundinella, whilst Miscanthus
is too near to Imperata to be rejected from Andropogonee.
Garnotia comprises about eight species from Hast India, China,
and Japan.

5. ARUNDINELLA, Raddi, includes Goldbachia, Trin., Acratherum,
Link, Thysanachne, Presl, and Brandtia, Kunth. It is the prin-
cipal genus of the tribe, and comprises about twenty-four species
spread over the tropical regions both of the New and the Old
World, but chiefly in Asia. It is generally adopted and fairly
characterized, though the habit and especially the inflorescence
vary much, the panicle being sometimes long, narrow, and dense,
or very large, loose, and spreading, with very numerous small or
minute spikelets, whilst in a few species it is short and dense,
forming almost an oval head with larger spikelets. The two
sections proposed by Nees—Meliosaccharum, with a small tooth
on the flowering glume on each side of the awn, and dcratherum,
in which the glume is quite entire, tapering into the awn—do not
prove to be well defined nor conformable to habit. A. flammida,
Trin., from Brazil and tropical Africa, has neither the habit nor
the character of the genus, but is in every respect a Trichopterya,
with which it was not compared by Nees, Trinius, or Doell,
because it was at first only known as Brazilian, and Trichopteryax
was supposed to be exclusively African.

6. Pumvosrerma, Munro, is a single Chinese species, nearly
allied to Arundinella ; but there are three lodicules to the flower
and no palea (unless one of the lodicules, although apparently in
the same whorl as the others, be really a small palea), and the
caryopsis is half exserted from the fruiting glumes as in some
species of Sporobolus. Phenosperma globosa, Munro, is a tall
grass with a very large loose panicle, the slender but rigid

60 MR. G. BENTHAM ON GRAMINEE.

branches distantly verticillate along the main rhachis. It was
first received from the Jardin des Plantes at Paris, where it had
been raised from seeds brought from China by the Pere David;
but it has since turned up among Shearer’s Kiu-Kiang plants.

7. Metinis, Beauv. (Zristegis, Nees, Suardia, Schrank), is a
single Brazilian species, Nees’s original type of the tribe. It is
very near Arwndinella, but remarkable for the long slender awn
of the third empty glume, whilst the flowering glume is short,
without any awn. Doell has reduced the genus to a section of
Panicum, a view in which I can by no means concur.

8. Triscenta, Griseb., is a single Cuban species unknown to
me, but from the author’s description it must be very near to the
following. 9. ArTHropogon, Nees, a single Brazilian species,
well described and figured by Kunth. So also is 10. Reynavpia,
Kunth, a single West-Indian species allied to Arthropogon; but
the awn, longest on the lowest glume, is gradually shortened and
reduced to a point on the flowering one, and there is no palea:
there are, however, four lodicules, a condition so unusual in
Graminex, that we might be tempted to consider the lowest pair
of lodicules, though close upon the others, as being in fact a
bipartite palea.

11. Ruyncautyrrum, Hochst., two or three tropical African
species, which appear to form a fairly distinct genus allied to Arun-
dinella, but approaching nearer to the Andropogonee in the long
hairs of the lower glumes. The generic name was originally Nees’s,
who applied it to a South-African plant of Drége’s, which proves to
be scarcely even a variety of the Panicum (Tricholena) roseum of
that country. Hochstetter and Steudel totally misunderstood
Nees’s genus when they added to it their R. grandiflorum and R.
ruficomum, which may now, however, retain those names, Nees’s
genus being suppressed.

12. Tuysanotana, Nees (Myriacheta, Zoll. and Mor.), is a single
tropical Asiatic species, a very tall grass with long broad leaves
and a very Jarge full panicle, with innumerable minute spikelets
in dense clusters along its long crowded branches. The flowering
glumes are more or less covered with rather long hairs; but these
hairs are so closely appressed and covered by the empty glumes
that Steudel could not see them, and published a supposed second
species as being destitute of them. ‘Trinius figured the plant as a
Panicum; by other early Indian botanists it was referred to
Agrostis.

MR, G. BENTHAM ON GRAMINEA. 61

13, CLErIstTacHyn, is a genus I have proposed for two plants,
one from East India, the other from tropical Atrica, which have
something of the aspect of Sorghum tropicum; but the spikelets
all hermaphrodite, and never in pairs, remove them from the
Andropogone to the Tristeginew. I purpose figuring the genus
in the forthcoming part of Hooker’s Icones.

Tribe V. ZoystEn.

I have composed this tribe of two groups or subtribes, which
might perhaps have been regarded as separate tribes, although the
difference between the two is only that which lies between the
Cenchrus group and Panicee proper. In the first group, Anthe-
phoree, the spikelets disarticulate from the rhachis of the inflo-
rescence or from the pedicels in little clusters of two to six, or
very rarely more; in the other group, or Zoysiee proper, the
spikelets are solitary, or very rarely two together on the pedicels.
In both groups the structure of the spikelets is generally that of
Andropogonex, sometimes slightly approaching that of Panicee,
but the pedicels are singly scattered or alternate along the inar-
ticulate rhachis of the spike or general inflorescence. The An-
thephorex have hitherto been usually placed in Panicew, as having
nearly the inflorescence of Cenchrus, but of which they have
not the hardened inner fruiting glume; the Zoysiex proper have
mostly been considered as Andropogonee, from which they differ
in inflorescence. Of the twelve following genera, the first six
belong to Anthephorex, the remaining six to Euzoysiee or
Zoysiew proper.

1. Hivania, H. B. K., in which I should include Pleuraphis of
Torrey, and, judging from the figure and description, Hevarrhena
of Pres], comprises five or six species dispersed over the Mexicano-
Texan region, extending into California. Although the forms
and proportions of the glumes of each spikelet vary much in the
different species, or even in different spikes of the same plant, the
genus as a whole is a natural one, and readily recognized by
each cluster consisting of three spikelets, the central one con-
taining a single fertile flower, either female or hermaphrodite, the
two lateral ones each with two male flowers. The spikelets are
often so closely sessile in the cluster, that it requires some care
to ascertain which glumes belong to each cluster, and the pairs
of male triandrous flowers of the lateral spikelets have sometimes
been described as single hexandrous flowers. The species I have

LINN. JOURN.—-BOTANY, VOL. XIX. a

62 MR. G. BENTHAM ON GRAMINER.

seen are H. cenchroides, H. B. K., H. Jamesii (Pleuraphis Jamesii,
Torr.), H. mutica (Pleuraphis mutica, Buckl.), H. sericea (Pleu-
raphis sericea, Nutt.), and a West-Texan species (Wright n. 758
and 2109, Berlandier n. 168, 1428) very near H. cenchroides, but
apparently distinct.

2. A.aorogon, Humb. and Bonpl. (Hymenothecium, Lag., Schel-
lingia, Steud.), extends in two species from Bolivia to Mexico.
The genus has at first sight much the aspect of the Asiatic
Melanocenchrus, or of some of the very short-spiked species of
Bouteloua, but the real affinity appears to be with Hilaria. The
spikelets usually vary from two to six in the cluster, mostly
with one hermaphrodite flower in each, though there are usually
one or two empty barren spikelets intermixed ; the clusters are
in a loose one-sided spike, each one very readily disarticulating
from its very short pedicel.

3. Caruestecuus of Presl, a single Mexican species, is only
known to me from his figure and description, which do not agree
with each other in some important particulars. He says that the
genus isalhed to Agopogon. Ihave no means of judging whether
that be really the case.

4, AntuEPHoRa, Schreb., is a very well-known and perfectly
characterized genus of five or six species, of which one is tropical
American, the others tropical or Southern African. Aypudeurus,
Hochst., quoted by A. Braun in ‘ Flora,’ 1841, p. 275, and by
some others, is Anthephora abissynica, Steud.

5. Tracuys, Pers., is a single well-known species from the East-
Indian peninsula, several times figured by the earlier botanists of
this century. It is slightly anomalous in the tribe by its spikes
being two together at the apex of the peduncle, and, as in An-
thephora, the excessive hardness of the clusters of spikelets after
flowering renders it difficult to trace their structure unless exa-
mined young. The name Zrachys was changed by Reichenbach
to Trachyozus, and by Dietrich to Trachystachys, as having been
preoccupied by zoologists, a plea not now regarded as sufficient.

6. Tracus, Hall. (Lappago, Schreb.), is a single annual very
well known as a common weed in tropical and temperate regions
almost all over the civilized world.

7. T.arrprs, Kunth, is a single tropical-African annual, extend-
ing eastward as far as Scinde, very well described and figured by
Kunth. It has been united by others with Tragus; but the
small spikelets, usually solitary or rarely two together on tho

ME. G. BENTHAM ON GRAMINEZ, 63

pedicel, and the very different shape and proportion of the glumes,
seem sufficient to maintain the genus as distinct.

8. Lornozspis, Dene., is a little slender East-Indian annual,
allied in some respects to Latipes, but with excessively minute
curiously shaped spikelets, so rapidly ripening and so very deci-
duous that it is very rare to find any on the specimens in an
examinable state. The plant was first sent home by Wallich
under the name of Holboellia, and was figured as such by Hooker
in the Botanical Magazine ; but in the meantime Wallich pub-
lished a Lardizabalous genus under that name in his Tentamen
of a Nepal Flora, and Decaisne therefore changed that of the
present grass to Lopholepis.

9. Nzunacung, R. Br., three Australian species, and 10. Pero-
vis, Ait. (Xystidium, Trin.), from the tropical regions of the Old
World, of which the species are variously estimated as from two to
seven, are both of them well-known genera, accurately described
and figured.

11. Lzprorurium, Kunth, founded on a specimen brought by
Humboldt from tropical America, is unknown to me. It is said
to be very near the Asiatic genus Zoysia, and, from the descrip-
tion, seems to differ chiefly in its widely distant geographical
station and in the presence of an additional lower empty glume.

12. Zoxsta, Willd. (Aatrella, Pers.), is a well-defined genus of
two or three maritime plants, dispersed over the shores of eastern
and southern Asia, Australia, and New Zealand, extending also to
the Mascarene Islands.

To these Zoysiew I have provisionally added a small Mexican
plant, the affinities of which are very puzzling, and which I
have described and figured as anew genus ScHarrnerRa, so named
after the collector from whom we have received it. At first sight
it seemed to bear some resemblance to Presl’s figure of Catheste-
chus ; but the structure of the spikelets is quite different, being
nearly that of Zoysia, whilst the general inflorescence, though on
a much smaller scale, approaches that of some species of Andro-
pogon (Cymbopogon) or of Apluda.

Tribe VI. AnDROPOGONED.
This tribe iy chiefly characterized by the spikelets in pairs at
each node of the articulate rhachis of the spike or of the branches
of the panicle, or in triplets at the end of each branch, and by

the inner glume under the fertile flower being much smaller and
G2

64: MR. G. BENTHAM ON GRAMINES.

thinner than the lower or outer empty ones, usually hyaline, and
often bearing a twisted or bent awn. The two spikelets of each
pair are either both of them perfect and fertile, or one of them
is male only or imperfect, or even quite rudimentary, and the
spikelets are often more or less surrounded by long silky hairs.
But to each of these characters there are exceptions in single
genera, which are retained in Andropogonee as agreeing with
them in most other respects.

The plants of this tribe are for the most part tropical or sub-
tropical, although a few are found in more temperate regions,
chiefly in the northern hemisphere. More than eighty genera
have at different times been proposed, which some botanists would
reduce to below twenty. Following as nearly as possible the
principles we have hitherto adopted, I have thought that the
following twenty-six may be admitted as fairly characterized,
referring them to four subordinate groups or subtribes—Sae-
charea, Arthraxea, Rottboelliee, and Andropogonee proper.

Saccharee comprise seven genera, in which the two spikelets
of each pair are homogamous, both of them hermaphrodite and
usually fertile, and the inflorescence paniculate, excepting Pogo-
natherum.

1. Imerrarta, Cyr., three or four species widely spread over the
tropical and subtropical regions both of the New and the Old
World, extending northwards to South Europe, China, and Japan.
In this and the following, Miscanthus, the branches of the panicle
are exceptionally inarticulate, showing an approach to the Tri-
steginee; but the long silky hairs and the very much reduced
hyaline flowering glume and palea retain them in Andropogonee.
Munro has shown that the common American J. caudata, Anders.,
is identical with the Old-World J. ramosa, Anders.; and I also
can find no difference between the two, any more than between
the American and the Old-World specimens of I. arundinacea.
Fournier has, however, proposed to separate the American forms
of the two species generically under the name of Syllepis, on the
plea of their having the two lodicules connate into a single large
truncate one, which I have in vain sought for in several different
American specimens. It is possible that Fournier may have
considered the small truncate palea as a pair of united lodicules,
but, if so, they are precisely the same in the Old-World species.

2. Miscanruus, Anders., as now limited, is a genus of eight
species, of which one is South-African, the others dispersed over

MR. G. BENTHAM ON GRAMINER. 65

Eastern Asia from the Malayan archipelago to Japan. It has
the inarticulate panicle-branches and most other characters of
Imperata, from which Andersson technically separated it by the
awn of the flowering glume. Exceptional unawned species occur
in so many genera where they are usually awned, that this can
scarcely be regarded as a generic character where there is nothing
else to separate the twoforms. Here, however, if we remove one
species from Imperata to Miscanthus, inflorescence supplies two
natural groups. In Imperata the panicle is long, narrow, and
dense, with short erect branches buried in the copious silky hairs,
the glumes are never awned, and there is only one, or rarely two,
stamens; in Miscanthus the panicle is loose, with long spreading
branches, the silky hairs are less dense and in one species almost
wanting, the flowering glume is in most species awned, and there
are always three stamens. The species known to Trinius were
by him included in Hulalia ; and Munro, whom I followed in
the ‘Flora Hongkongensis,’ restricted the name Eulalia to the
species now constituting Miscanthus; but as the true Hulalia of
Kunth is the type of a section of the very different genus Pol-
linia, I have thought it necessary to adopt Andersson’s later name
Miscanthus. Besides his species, I would include in the genus
MM. fuseus (Eriochrysis fusca, Trin., H. attenuata, Nees) from Hast
India, WM. saccharifer (Imperata saccharifera, Anders.), from North
China, which has the inflorescence and stamens, but not the awns,
of the other species, and IL. cotulifera (Hulalia cotulifera, Munro)
from Japan, which has scarcely any of the hairs of the other species.
Steudel proposed the latter as a distinct genus under the name
of Eccoilopus.

With 8. Saccuarum and 4. Erranraus commence the series
of true Andropogonew with the branches of the panicle arti-
culate ; and these two genera are so closely connected that they
might well be reunited, although they are now almost universally
recognized as distinct. There might indeed be no great objec-
tion to consider both, as well as Pollinia and Spodiopogon, as sec-
tions of one large genus. As now limited, Saecharwm is chiefly
characterized by the compound panicle, usually dense, sometimes
very large, and the spikelets very small without any points or
awns to the glumes. The species are supposed to be about ten,
the typical ones belonging to the tropical or subtropical regions of
the Old World, amongst which the well-known sugar-caue is now
extensively cultivated also in America, The genus would also

66 MR. G. BENTHAM ON GRAMINEZ.

include 8. Nareya (Eriochrysis Nareya, Nees) and 8. longifolia,
Munro (Eriochrysis longifolia, Munro), from East India, 8. pallida
(Eriochrysis pallida, Munro) from South Africa, and S. cayennense,
the typical Hriochrysis of Beauvois, which last differs only in the
very dense almost spike-like panicle.

4, Exrantuvs, Mich. (Ripidium, Trin.), would be a more satis-
factory genus if it could be restricted to the two old species
E. saccharotdes, Mich., from North America, and Z. Ravenna,
Beauv., from the Old World; but besides the above-mentioned
connexion with Saccharum, there are several South-American
species which run very closely into Pollinia. On the whole, it
seems best to consider Hrianthus as an intermediate genus
between Saccharum and Pollinia, having the inflorescence of the
former, but the flowering-glume more developed into a point or
awn almost as in Pollinia. It would then consist of about twelve
species, amongst which #. stricta, Nees, from North America, hag
no hairs on the rhachilla, but only a short pubescence on the
glumes.

5. Spopiorocon, Trin., differs from Pollinia, as Chrysopogon
does from Andropogon, chiefly in inflorescence. The short
branches of the panicle bear three spikelets, one sessile between
two pedicellate, and occasionally there is a pair of spikelets below
the three terminal ones ; but the branches never form the regular
spikes of Pollinia. Besides the original S. sibiricus, Trin., we
have two additional species, S. pogonanthus, Boiss., from the
Levant, and S. albidus (Andropogon albidus, Wall. Cat. Herb. Ind.
n. 8821), from Hast India. The generic name has also been often
misapplied. S. angustifolius, Trin., is a Pollinia; some others of
his species with 2-flowered spikelets belong to Ischemum. Four-
nier’s Mexican Spodiopogons are evidently species of Hrianthus ;
his 8. foliata indeed (Bourgeau, n. 2979) appears to me to be
identical with the original LZ. saccharoides, Mich.

6. Porrinta, Trin., is now a genus of about twenty-five tropical
or subtropical Old- World species, with the inflorescence of the sec-
tion Gymnandropogon of Andropogon, and the homogamous spike-
lets of Saccharum and Hrianthus; the spikelets are in pairs along
the simple branches of the panicle; these branches either few,
almost digitate at the end of the peduncle, or more numerous and
scattered along the main rhachis. The genus is divisible into
two very natural sections :—1. Eulalia, with the spikes and pedicels
covered with long silky or rufous hairs as in Zrianthus, includes

MR. G, BENTHAM ON GRAMINED. 67

P. aurea, Benth. (the original genus Hulalia, Kunth), P. articu-
lata, Trin. (Pogonatherum contortum, Brongn.), P. eriopoda, Hance
(Spodiopogon angustifolius, Trin.), P. longisetus (Erianthus lon-
gisetus, Anders.), P. versicolor (Erianthus versicolor, Nees), P.fili-
folius (Lrianthus filifolius, Nees), and a few others. 2. Lepta-
therum, with slender spike-like branches, of which the hairs are
few or short, so as to appear sometimes quite glabrous; this
section includes P. glabrata, Trin. (Hulalia glabrata, Brongn.),
P. nuda, Trin. (P. imberbis, Nees), P. Willdenowianum (the genus
Microstegium, Nees, P. lancea, Nees, published also by Nees as
his genus Leptatherwm, and probably also Steudel’s Nemastachys).
Sprengel’s Pollinia would have had the right of priority over
Trinius’s; but that proved a farrago made up of a few heteroge-
nous species of Andropogon, Chrysopogon, and Pollinia.

7. PogonaTHERUM, Beauv. (Homoplitis, Trin.), is a single tro-
pical and subtropical Asiatic species, very well marked by its
slender, much branched habit, the single spikes, and the slender
awns arising as well from the second empty glume as from the
flowering one.

The Arthravee, or second group of Andropogonesx, consist of
three genera, which have the inflorescence of Pollinia; but the
second spikelet of each pair is generally reduced to a bare stipes,
or ig even quite deficient, bringing a few species very near to the
Zoysiew, differing chiefly in their subdigitate spikes, whilst a few
others, in which the spikes are single, have the rudiment of the
second spikelet of true Andropogones.

8. Arocoris, Nees (Amblyachyrum, Hochst.), has five or six
species from Hast India or the Malayan archipelago, characterized
by the very broad truncate outer glume enclosing the rest of the
spikelet. Among the species 4d. Royleanus, Nees (Ischemum
paleaceum, Trin., Andropogon paleaceum and A. himalayensis,
Steud.), is remarkable for the awn often (but not always) reduced
toasmall fine point, or even entirely wanting ; and A. tridentata
(Andropogon tridentatus, Royle) has, on the contrary, a very long
awn, and the young spikes are usually enclosed in a large spathe-
like bract.

9. Dimerta, R. Br. (Haplachne, Presl, Didactylon, Zoll. and
Mor., Psilostachys, Steud., Pterygostachywm, Nees), about ten
species from the Indo-Australian region, has very slender spikes,
the lower empty glumes very narrow and rather rigid, and usually,

if not always, only two stamens,

68 MR. G. BENTHAM ON GRAMINES.

10. AnTuraxon, Beauv. (Pleuroplitis, Trin., Batratherum,
Nees, Lucea, Kunth, Lasiolytrum, Steud., Alectoridia, A. Rich.,
Psilopogon, Hochst.), has also about ten species, chiefly from
the Indo-Australian region, but extending on the one hand to
China and Japan, and on the other to tropical Africa. The
spikes are slender as in Dimeria; but there are three stamens,
and the lower empty glume is broad but acute, not truncate as in
Apocopis.

Rottboelliew, the third group of Andropogonex, is often re-
garded as a distinct tribe, characterized by the simple spike,
the spikelets in pairs at each notch or excavation of the rhachis,
the one sessile, the other pedicellate, and no awn to the flowering-
glume. There are, however, as in other subtribes, here and
there exceptions to one or more of these characters. We have
seven genera.

11. Etronurvs, Humb. and Bonpl., has about twelve species,
chiefly South-American or African, with, however, one Austra-
lian and one from the East-Mediterranean region. They all
differ little from Rottboellia besides the long silky hairs which
clothe the spike, thus connecting Rottboelliee with other An-
dropogoner. LE. hirsuta, Munro (Rottboellia hirsuta, Vahl), has
been proposed by Boissier asa distinct genus Lasiurus, as having
the spikelets in threes instead of in twos at each node of the
rhachis. But that character is by no means constant; in several
specimens I have found the spikelets in threes or even in fours
at the lower nodes; but in others they are in the normal pairs
from the base of the spike.

12. Rorrsortiz1a, Linn. f.,a tropical or subtropical genus widely
spread, but chiefly in the Old World, has been either extended to
nearly the whole subtribe or very variously restricted to a small
number or to asingle species. It seems best characterized by in-
cluding all those which have the simple terete spike, without the
hairs of Hlionurus or the peculiarities of the four following genera.
It would contain about eighteen species, amongst which several have
been proposedasmonotypic genera. Calorhachis, Brongn.,is RB. mu-
ricata, Retz (R. glandulosa, Trin.); Peltophorus, Desv.,is R.myurus:
in both of these the lowest or outer glume of the perfect spikelet is
rigid and bordered on each side at the apex by a membranous wing,
which, however, is also present, but much less prominent, in R.
rugosa, Nutt. Phacelura, Griseb. (Pholiurus, Trin. in Spreng. Neue
Entd. ii. 67, not of the Fundam. Agrost.), is the Oriental R. digi-

MR. G. BENTHAM ON GRAMINES. 69

tata, Sibth. (R. Sandorii, Friwaldsk.), a species striking for the
long spikes, occasionally though very rarely branched at the base,
and from the rather large spikelets with acuminate outer glumes
showing an approach to some Vossie, but scarcely sufficiently
distinct from Rottboellia to be kept up as an independent genus.
Cymbachne, Retz, a Bengal grass, has been referred by Willdenow
to Rottboellia, Retz’s character does not quite agree; but the
plant has not since been identified, and must remain doubtful.
Apogonia, Fourn., comprises two Mexican species which I am
unable to distinguish from Rottboellia : Nuttall’s section Apogonia
of Rottboellia is a species of Elionurus, very closely allied to, if
not a variety of, #. ciliaris, H. B. K.

13. Opniurus. This genus, as first proposed by Gartner, in-
cluded two very different plants separated by Brown as Lepturus
and Ophiurus. As the latter is now limited, it differs from Rott-
boellia only in the absence of the second sterile spikelet of each
node, at least in the upper part of the spike or inflorescence. It
consists of three, or perhaps four, Asiatic, African, or Australian
species :—O. corymbosa, Gertn. (O. ethiopica, Steud.), O. mono-
stachya, Presl (O. undulata, Nees), and O. levis (Rottboellia levis,
Retz, R. perforata, Roxb.). The latter species is remarkable for
having the spikelets in the lower part of the inflorescence in pairs
at each node as in Rottboellia, but the two of each pair separated
by a kind of partition dividing the cavity of the rhachis into two ;
it has therefore been raised to a genus by Kunth as Mnesithea
and by Nees as Thyridostachyum. Generally, however, in the
upper part, and sometimes in the whole inflorescence, the sterile
spikelet is wanting, as in Ophiurus, especially in the young
spike, for the upper or Ophiurus portion appears to fall away
very readily, leaving only the Mnesithea part persistent. Lepturus,
Br., is now classed in the tribe Hordeez.

14, Rarzesurata, Kunth, a very elegant little flat-spiked
Burmese grass, and 15. Manisuris, Linn., a common tropical
weed with little globular spikelets, have both been well described
and figured.

16. Hemartunrta, Br., contains two or three tropical weeds or
maritime grasses, separated from Rottboellia chiefly on account of
the flattened and less distinctly articulated rhachis of the spike,
and the curious way in which the stipes of the sterile spikelet is
adnate to the rhachis, so as to make it appear sessile and almost
opposite to a fertile spikelet, which really belongs to the next

70 MR. G. BENTHAM ON GRAMINER.

superior node. These characters, though generally well marked,
are sometimes more or less obscure.

17. Vossta, Wall. and Griff., closely connects the Rottboellies
with Ischemum. As in the former, the flowering glume is always
unawned, and the rhachis of the spike is rigid and deeply notched,
but the lower empty glume, at least of the pedicellate spikelet, is
produced into a long point or awn; there are generally several
spikes or simple branches along the common peduncle, and there
is in each sessile spikelet a male flower below the terminal fertile
one, as in Ischemum. The genus was originally established on a
handsome semiaquatic East-Indian grass, which has since been
found also in tropical Africa, and two or three additional species
have reached us from the same country. We should also refer to
Vossia the Ischemum speciosum of Nees from East India. Hremo-
chloa, a Japanese plant described by Biise, is unknown to me; but
the character given, if I correctly understand it, agrees well with
that of Vossia.

To the fourth group, or subtribe Huandropogonee, may be
referred nine genera, in which the two spikelets of each pair are
heterogamous and the flowering glume of the fertile one is more
or less awned ; and in the first five the spikelets are in many pairs
along the rhachis of the simple spikes or panicle-branches. These
nine genera are increased to twenty-one by Andersson and others,
whilst Steudel unites seven out of the nine under his Andropogon.

18. THELEPogon, Roth (Jardinia, Steud.), comprises one East-
Indian and two or three tropical-African species, all very elegant
and closely resembling each other. Their inflorescence is that of
Vossia, whilst the spikelets are nearer those of Ischemum, but re-
markable for the rigid tuberculate outer empty glumes. Nees, in
working up Wight and Arnott’s Peninsular grasses, gave Roth’s
name to a very different grass (Ischemum semisagittatum, Roxb.),
adding the observation that Roth’s ‘description is very bad. The
fact is, however, that it is Nees who was mistaken in his identi-
fication, whilst Roth’s description of the true plant is excellent.

19. Iscumuum, Linn., as now understood, has about thirty
species, widely dispersed over the warmer regions both of the New
and the Old World, the chief character connecting them being
that the sessile spikelets have a male flower below the terminal
fertile one. The spikes are also usually stouter than in Andro-
pogon, and the genus is a fairly natural one. Beauvois restricted it
to the £. muticum, Linn., in which the awn of the flowering glume

MR. G. BENTHAM ON GRAMINER. 71

is small and hair-like or sometimes entirely wanting, and proposed
a genus Meoschium, adopted by Nees, for the other species in which
the awn is more developed. Trinius considered as true Ischema
only those in which the pedicellate spikelet has only a male flower
or empty glume, and added those in which that spikelet has two
male flowers to his genus Spodiopogon, notwithstanding the dif-
ference in inflorescence &c. Ischemopogon, Griseb., is I. latifolium,
Kunth, and Hologamium, Nees, is I. laxwm, Br., both species
with two-flowered pedicellate spikelets, as is also the case in J.
insculptum, Hochst., and I. macrostachyum, A. Rich., from tropical
Africa, and probably also in Forsk&hl’s genus Sehima, of which
we have no authentic specimen. J. pectinatum, Trin., I. leersioides,
Munro, L£. ophiuroides, Munro, with a fourth unpublished species,
all from tropical Asia, form a distinct section (Pectinaria), with
slender elegant simple spikes, and the larger glume of the sessile
spikelets pectinate-ciliate.

20. Tracuyrodon, Nees, as limited by Andersson, and 21.
Hereroroaon, Pers., closely resemble each other in their simple
spikes with appressed imbricate spikelets and long rigid twisted
awns ; but in Zrachypogon the sessile spikelet of each pair is male
or sterile and unawned, and the pedicellate one fertile and awned,
whilst in Heteropogon the sessile one is fertile and awned, and
the pedicellate one male or sterile and unawned. Andersson enu-
merates eleven species of Zrachypogon, one from South Africa, the
others from tropical or subtropical America; but several of the latter
can scarcely be regarded as more than slight varieties. Of Hetero-
pogon there are two well-marked species, H. contortus, Roem. and
Schult. (A. hirtus, Pers.), now very common in most warm regions
and extending to the Mediterranean region and to North America,
and H, melanocarpus, Ell. (H. Roylei, Nees, H. acuminatus, Trin.,
Trachypogon scrobiculatus, Nees), which is in North and South
America as well as in East India. Besides these, three or four
South-African species have been referred to Heteropogon, but are
of somewhat doubtful affinity.

22, AnpRorogoy, Linn., taking it within the limits assigned to
it by Munro, including all the species of the subtribe with spike-
like simple branches to the inflorescence, and without the pecu-
liarities of the three preceding genera, is still a somewhat poly-
morphous genus of perhaps a hundred species, very abundant
within the tropics, but well represented also in Europe, temperate
Asia, North America, South Africa, and Australia. The fourteen

72, MRE. G. BENTHAM ON GRAMINER.

genera into which it has been divided may be fairly reduced to
the following five, perhaps too artificial, sections:—1. Schizachy-
vium, about a dozen species, with the spikes always single upon
each peduncle. The genus Schizachyrium, Nees, was limited to
a few species in which the spike is slender and not very hairy.
Diectomis, H. B. K., is the American A. fastigiatus, Sw., found
also in tropical Africa, which has a more rigid spike and the
second empty glume conspicuously awned. Homceatherum, Nees,
is an Asiatic species scarcely to be distinguished from the same
A. fastigiatus. In 2. Cymbopogon, the spikes, often very silky-hairy
or woolly, are in pairs on each peduncle, and the peduncle partly
or wholly enclosed in the sheath of a leafy or spathe-like bract.
The species are numerous, chiefly in the Old World, and include
the lemon-grass and its allies. Andersson has divided the section
into two genera, Gymnanthelia and Hyparrhenia, and perhaps
more; but as he has never published their characters, I am unable
to form any clear idea of them. It would appear, however, from
the species quoted, that A. schaenanthus and its allies would be-
long to Gymnanthelia, and A. hirtus and its allies to Hyparrhenia.
3. Gymnandropogon, has two or more spikes sessile at the end of the
peduncle, without any sheathing-bract. The species are nearly as
numerous as those of Cymbopogon. Amongst them, A. annulatus,
Forsk., though closely allied to the common A. Ischemum, forms
the proposed genus Dichanthium, Willem.; A. serratus, Retz,
with a broad herbaceous outer glume, is Trinius’s genus Lepeo-
cercis ; and it is most probable that Steudel’s Huklastaxon is the
common American A. virginicus. 4. Amphilophis, Trin., would
include A. laguroides, DC., and A. argenteus, DC., from tropical
America, with A. scandens, Roxb., and A. Vachellii, Nees, from
tropical Asia, and a few others, differing from Gymnandropogon in
the more numerous, usually long and often pedicellate spikes,
sometimes even divided at the base, forming almost a saccharoid
panicle. 5. Vetiveria, Thou. (Mandelorna, Steud.), is the well-
known Vitiver, A. muricata, Retz, to which Munro would redu e
as varieties A. nigritana, Benth., and Vetiveria arundinacea,
Griseb., a species frequent in Hast India and tropical Africa and in-
troduced into America, distinguished by its numerous spikes verti-
cillate along the axis of a long simple panicle, all glabrous or only
minutely hairy, and the awn of the flowering glume often very
much reduced. Beauvois’s genus Anatherwm, sometimes supposed
to be specially destined for this plant, included also all the species

MR. G. BENTHAM ON GRAMINED. 73

of the sections Cymbopogon and Gymnandropogon, in which the
awn is much reduced or obsolete. Ageniwm, Nees, from his cha-
racter, would also refer to one of these species without prominent
awns.

23. Curysopogon, Trin. (Rhaphis, Lour., Centrophorum, Trin.),
and 24. Soranum, Pers. (Blumenbachia, Koel.), are two genera
very nearly allied to each other and differing from Andropogon, as
Spodiopogon does from Pollinia, chiefly in their inflorescence;
the branches of the panicle bear three spikelets at the end, a
sessile one between two pedicellate ones, and occasionally only
one or two pairs below on the same branch. They were both
included by Linnzus, and afterwards by Brown, in Holcus, a
name since restricted to that portion of the old genus which
belongs to Avenacer. Chrysopogon, as now constituted, has nearly
twenty species, chiefly tropical or subtropical, but including also
the European C. Gryllus and some other temperate species. The
genus may be divided into two natural sections: in the typical
form the pedicellate spikelets usually contain a male flower; in
the section Stipoides, exclusively American, it is reduced toa long
hairy stipes rarely bearing a minute rudimentary glume. This
section includes C. nutans, C. avenaceus, C. stipoides, C. Minarum,
and a few others. Sorghum differs from Ohrysopogon in habit,
in the scarcely articulate branches of the panicle, and in the
glumes of the fertile spikelets more hardened after flowering.
The number of species is very uncertain, for, of the two prin-
cipal ones, S. halepense is so widely spread as a tropical or sub-
tropical weed, and S. vulgare so long and so generally cultivated
in warm regions for a variety of purposes, as to have produced
a great variety of forms, raised by many to the rank of species.

25. Awrutstrria, Linn. fil. (Lhemeda, Forsk.), if taken as a
whole, is a very natural genus, of about a dozen species from
the warmer regions of the Old World, easily recognized by its
inflorescence. The spikelets are in short dense spikes or clusters,
usually seven together, of which the four lower ones (two pairs)
are either empty or with a male flower in each, and are placed
apparently in a whorl, forming a kind of involucre round the
three inner ones, which, as in Chrysopogon, are one sessile between
two pedicellate ones. In a few species the number of spikelets
ig raised to nine, or even to eleven, by the intervention of one or
even two pairs of spikelets between the involucral and the ter-
minal ones. These slight differences in the number or in the

74 MR. G. BENTHAM ON GRAMINEZ.

pedicellation of the spikelets have induced the proposal of distinct
genera for most of the species, and several of them have been
adopted by Andersson in a monograph most carefully worked up
in as far as the materials at his command admitted, but in which,
for want of access to a sufficiently rich library, he is much mis-
taken as to several of the synonyms quoted. These proposed
genera are:—1. Aristaria, Jungh., for A. frondosa, Br. (A. Jung-
huhniana, Nees), which forms the section Heterelytron of Anders-
son, but not Junghuhn’s genus of that name. 2. Perobachne,
Presl, is 4. arundinacea, Roxb., forming Andersson’s subsection
Chrysanthistiria. 3. Andersson’s subsection Euanthistiria for the
common A. ciliata, Linn., and its immediate allies, to which some
botanists would restrict the genus. Andersson distinguishes twelve
species, adding at the same time that they might well all be re-
duced to varieties of a single widely-spread species. 4. Andro-
scepia, Brongn. (Heterelytron, Jungh.), was founded originally on
the A. gigantea, Cav., but became a very unnatural group when
made to include A. (Androscepia) anathera, Anders., which very
closely resembles A. (Zuanthistiria) minuta, Anders., and a variety
armata, Anders., of A. gigantea, which is much nearer to the A.
(Perobachne) arundinacea. 5. Iseilema, Anders., containing two
East-Indian and one Australian species, and 6. Hwotheca, Anders.,
comprising 4. abyssinica, Hochst., from tropical Africa, and A.
fasciculata, Thw., from Ceylon, have each a peculiar habit and
characters, sufficient to maintain them as sections. 7. Germainia,
Balansa, has, perhaps, two closely allied species—A. caudata, Nees,
from Khasiya and China, and the typical A. capitata fron Saigou;
the latter, however, which I only know from Balansa’s figure and
description, is exactly like the Chinese plant, except that there
appear to be rather more spikelets in the cluster.

26. Aptupa, Linn., is now universally recognized as a distinct
and natural genus, limited to the two tropical-Asiatic species ori-
ginally assigned to it by Linnaeus, though his character was even
then very imperfect, and rendered still more so by the subsequent
addition of the very different American Zeugites, which Schreber
afterwards restored as anindependent genus. Beauvois, however,
threw every thing into confusion; for it is evident from his figures
that his Diectomis is A. aristata, Linn., and his Calamina is A,
mutica, Linn., though in drawing up his character for the latter
he combined it with some species of Anthistiria. Beauvois’s
Apluda is certainly different, probably a Chrysopogon.

MR. G. BENTHAM ON GRAMINE. 75

Series B. POACEZ.

Having already explained the difference between the two
primary divisions of Graminee, I need only repeat here that
the main characters of Poacex consist, firstly, in the want of any
articulation of the pedicel below the lower empty glumes, which
remain persistent after the fruiting one has fallen away, or fall
away separately, and, secondly, in the male or imperfect or rudi-
mentary flowers, when present, being above, not below, the fertile
one. The former character is all but universal; but from the
latter one exceptions are not very rare, besides that, where there
is only one flower without any continuation of the rhachilla
beyond it, the character entirely fails. I should add that in
some tribes of Poace there are two or more perfect flowers in the
spikelet, which is not the case in Panicacer ; and may now pro-
ceed to examine in detail the eight tribes into which this second
series may be divided.

Tribe VII. PHaLanripEn.

The close affinity of this tribe and the Oryzee has been
generally admitted, and the two are usually placed in juxtapo-
sition; I had even proposed their consolidation into a single one
in the ‘Flora Australiensis,’ They have in common the im-
portant character of the scale immediately under the single
perfect terminal flower being keeled or one-nerved, so as to make
it a matter of discussion whether it be a glume terminal on the
main axis or rhachilla of the spikelet, or a palea at the base of a
secondary floral axis. The deciduous part of the spikelet of Phala-
ride with its four glumes(or three glumes and a palea) is precisely
as in Oryzew ; but there are in addition, below the articulation,
the two persistent empty glumes characteristic of Poacee. The
spikelet, therefore, in this tribe consists of six glumes (or five and
a palea), the lowest pair empty below the articulation; the second
pair, above the articulation, corresponding to the lowest two glumes
of Oryzew, are usually empty and small, sometimes reduced to a
small bristle, rarely enclosing each a small palea or a male flower;
the upper pair (or glume and similar palea) enclosing the terminal
fertile flower and fruit, without any continuation of the rhachilla
above it. A slight apparent exception will be mentioned under
Phalaris itself; and in the genus Cinna of Agrostidee and a very
few Bambusee the palea of the fertile flower is, at least apparently,

76 MR. G. BENTHAM ON GRAMINES.

one-nerved, but otherwise the character of Phalaridew is constant.
They comprise the following six genera :—

1, Enruarra, Thunb. (Trochera, L. C. Rich.), has twenty-four
species, of which two are from New Zealand, two from the Mascarene
Islands, and all the rest from South Africa. In them the glumes
of the second pair are the largest, empty and usually awned, and
the fertile flower has six stamens. 2. Micronana, Br., including
Diplax, Hook. f., has five Australian or New-Zealand species,
differing from Hhrharta only in the number of stamens reduced
to four or two. 3. Trrrarruena, Br., four Australian species,
with four stamens to the flower as in Mcrolena, but the glumes
are in less regular pairs, all unawned, and the fourth (one of the
second pair) alone the largest. The panicle is also almost always
contracted into a spike, not, however, so dense and cylindrical as
in the following two genera.

4. Puatantis, Linn., has nine or ten extratropical species,
chiefly from the Mediterranean region, but also extending to
North and South America. In this genus it is the lowest two
persistent empty glumes that are the largest, usually very flat,
and often winged on the keel, the second pair (like the lowest in
Oryza) very narrow, sometimes reduced to small bristles, those
of the upper pair thin and hyaline; avd sometimes in both of
them, but almost always in the uppermost one, the central nerve
is very faint or quite obsolete, a character adduced as an argu-
ment that this upper one is a two-nerved palea on the floral axis,
and nota glume on the main rhachilla. The two nerves are,
however, very faint, and the central keel is usually marked by a
line of hairs on the outside, and the question remains a moot one.
In the majority of species the panicle is contracted into a dense
globular or cylindrical head; but in P. arundinacea, Linn., a stout
tall species, forming the genus Digraphis, Trin. (Baldingera, Gertn.,
Meg., and Schrad., T'yphoides, Mcench), the inflorescence, though
still very dense, is more or less branched or interrupted. This
genus has also been supposed to be distinguished by the want of
the broad wings of the outer glumes, so conspicuous in the com-
mon P. canariensis ; but these wings are very narrow in P. para-
doxa, Linn., and entirely disappear in P. intermedia, Bosc (P.
americana, Hil.) , leaving no available character to separate Digraphis
generically.

5. AntHoxantHUM, Linn., has four or five European species,
of which one is now widely spread over various regions of the

MRE. G. BENTHAM ON GRAMINEA. "9

globe, but often only as an mtroduced weed. One at least of the
glumes of the lowest pair is the largest of the spikelet, as in
Phalaris; those of the second pair, though small and without
flowers, have a dorsal awn. The panicle is usually cylindrical
and spikelike.

6. HrzrocHtoa, Gmel.(Savastana, Schrank, Disarrenuwm, Labill.,
Torresia, Ruiz and Pav.), about eight species from the colder or
mountain regions both of the northern and the southern hemi-
spheres, is usually referred to Avenacem next to Holcus; but it
appears to me to be much nearer to Anthowanthum, from which
it differs in its looser paniculate inflorescence, and in the glumes
of the second pair being but little smaller than the lower ones,
and frequently, but not always, enclosing each a male flower.
Ataxia, Br., one or two Asiatic and two South-African species,
forms a section of Hierochioa, differing slightly from the typical
form in the glumes of each pair being more unequal, the lower
one only of the second pair (rarely both) having a male Aower.
A. mewicana, Rupr., seems to connect the two sections.

Tribe VIII. Agrosten.

The large tribe Agrostee is one of the most difficult to cireum-
scribe satisfactorily, or to divide into definite genera. We have
taken it nearly in the sense given to it by Trinius, so as to in-
clude the Stipex, of which other botanists make a distinct tribe ;
and we have adopted thirty-seven genera, a number which some
would extend to above eighty, whilst others might reduce it to
about thirty. Their general character is to have a single flower
in each spikelet, either apparently terminal as in Panicacee, or
with a slight bristle-like continuation of the rhachilla beyond it;
and from these Panicacee they are constantly distinguished by
the pair of empty glumes persistent below the articulation of the
rhachilla, without any empty glume or male flower intervening
between the articulation and the flowering glume. The single
flower in the spikelet, which separates the tribe from the follow-
ing ones, is not so positive a character, as it occurs also in one
genus of Avenew, in a few genera of Chlorides, and occasionally
in a few exceptional species of some genera of Festucee, which
cannot well, from inflorescence or other accessory characters, be
included in Agrostew. There are also two species of Sporobolus
which approach the Isanthes in having frequently two flowers ;
and in Coleanthus the lower empty glumes are entirely deficient.

LINN. JOURN.—BOTANY, VOL. XIX. H

78 MR. G. BENTHAM ON GRAMINEZ.

Trinius, in his elaborate monograph of the tribe, divided it into
three primary groups or subtribes—Vilfez with the callus scarcely
prominent or quite obsolete, Agroste with the callus globular,
and Stipex with the callus obconical. In this I feel unable to
follow him. In the first place he does not appear to have con-
sidered what the so-called callus really is. It is not, as the name
would suggest, an appendage to the base of the flowering glume,
or, as he would have termed it, to the flower, but only the upper
or principal part of the rhachilla or axis of the spikelet, to which
the glume and its enclosed flower are attached, and which breaks
off immediately above the persistent empty glumes. Its shape
depends on the distance at which the flowering glume is attached
above the empty ones, a distance very variable throughout the
Order. And although the long or the short interval may be
more prevalent or even constant in some genera, yet I have never
found the variations so precise as to be defined by actual measure-
ment, and the species are numerous, even in Stipa itself, where
it is doubtful whether we should call it long or short. It is
sometimes a useful accessory character, but, I believe, never
positive enough to be regarded as subtribual. It is true that no
other simple absolute character has yet been proposed for the
subdivision of the tribe; but we are obliged, here as elsewhere,
to take a combination of characters, to each of which an occa-
sional exception must be allowed. Acting on this principle, we
might, whilst following in many respects the arrangements of
Kunth and others, admit thirty-seven genera of Agrostex, dis-
tributed in four fairly natural subtribes, all four of wide geogra-
phical range, but chiefly in temperate regions, the tropical species
mostly confined to mountain districts, and no genus, except a few
monotypic ones, exclusively tropical.

Our first subtribe, StipEm, is the long-established one of that
name, slightly extended so as to include Oryzopsis, Muehlen-
bergia, and their immediate allies, the close connexion of which
with Stipa has been frequently suggested. The subtribe thus
formed would be characterized by the paniculate inflorescence
not condensed into the cylindrical spike of Phleoidex, by the
rhachilla of the spikelet not produced beyond the flower except
in the single species of Brachyelytrum, by the awn of the flowering
glume termiual, not dorsal as in Euagrostex, and especially by
the grain being very closely enveloped in the fruiting glume.
In the majority of species these characters are well marked; but

MR. G. BENTHAM ON GRAMINEA, 79

in the larger genera there occur occasional exceptions more or
less decided, which prevent our taking any single one of them as
an absolute test. The subtribe would include the following eight
genera, in the first five of which the fruiting glume is more or
less hardened or rigid as in Panices; in the succeeding three it
is thinner, though still closely pressed on the grain.

1. Anistipa, Linn., is now a genus of at least a hundred
species, abundant in all the warmer regions of the globe, but also
represented by a few species in Europe and temperate Asia, and
by several in North America. With few exceptions it is most
readily recognized by the long, fine, three-branched awns, the
lateral branches opposite and spreading. Doell adds to the
generic character three lodicules as in Stipa; and Nees describes
three lodicules in some South-African species; but all other
Agrostologists describe two only, and I have never found more
than that number. It is probable that both Nees and Doell
mistook for the third lodicule the palea, which in many species
is very thin and scarcely, if at all, larger than the lodicules. The
genus is divided into three fairly marked sections, which Beau-
vois, Necs, and some others have raised to the rank of genera.
In (1) Chetaria, Beauv., the flowering glume is continuous with
the awn without any articulation, and though much longer than
the empty glumes, and often much attenuated at the end, is
neither quite awn-like nor decidedly twisted below the branches.
Amongst its species, Curtopogon was proposed as a genus by
Beauvois for the North-American A. dichotoma, Mich., in which
the lateral branches of the awn, instead of diverging from the
central one, are short and erect at its sides, showing more or less
distinctly that they are continuations of the lateral nerves of the
glume. It is probable that this is the case throughout the genus,
only that the lateral nerves before they diverge are so closely con-
solidated with the central one as to be undistinguishable from it.
The genus Ortachne was proposed by Nees for two or three
Mexican or Columbian plants, originally published by Kunth as
species of Streptachne, Br., and afterwards transferred by him to
Aristida, in which the lateral branches of the awn are very short,
sometimes minute or even quite obsolete, thus nearly connecting
the section Chetaria of Aristida with the section Aristella of
Stipa, but in the narrow base of the rhachilla, and some other
minor points, nearer to the former than to the latter. Ortachne
retorta, Nees (in Steud. Gram.), is probably a true Stipa. In

m2

80 MR. G. BENTHAM ON GRAMINER.

(2) Arthratherum, Nees, the awn is decidedly articulate on the
glume and much twisted above the articulation below the branches,
the flowering glume itself much shorter than the lower empty
glumes, instead of exceeding them as in Chetaria. In (3) St-
pagrostis, Nees, the awn is articulate on the glume, as in Arthra-
therum, but scarcely twisted, and above the branches elegantly
plumose, the branches also being plumose in some species; whilst
in others, forming Figari and De Notaris’s proposed genus Schis-
tachne, the central awn alone is plumose, the lateral branches
short and glabrous. All, however, are most conveniently in-
cluded in the great genus Aristida.

2. Stipa, Linn., is almost as numerous and as widely spread
as Aristida. It is also strongly characterized, as to the great
majority of species, by the narrow, rather hard fruiting glume,
carrying off a rather long or obconical internode of the rhachilla
(or so-called callus), by the long undivided awn more or less
articulate on the glume and usually twisted at the base, and by
the presence of three lodicules ; but the exceptions to one or more
of these characters are more numerous than in Aristida; the in-
ternode of the rbachilla varies much in length and in shape, the
articulation and twist of the awn gradually disappear in some
species, and the third lodicule, though often as large as the others,
is sometimes much smaller or even quite obsolete. The genus is
also not so clearly divisible into sections as Aristida, although
several genera have been proposed for more or less aberrant
species. acrochloa, Kunth, includes 8. tenacissima, Linn., and
8. arenaria, Brot., both from the Mediterranean region, remark-
able for their large membranous glumes, the flowering one shortly
bifid at the apex. In dAristella, Bertol., founded on SV. aristella,
Linn., a European and Mediterranean species, in Streptachne, Br.,
a single Australian species, and in Orthoraphium, Nees, two or
perbaps three East-Indian species, the flowering glume is 2-toothed
or shortly bifid at the apex, the awn scarcely or not at all articu-
late, and the internode of the rhachilla very short, though still
perhaps slightly thickened under the flowering glume. The
S. aristella, however, is very closely connected with typical Stipe
through S. sibirica, Lam., 8. Redowskii, Trin., and 8. altaica,
Ledeb. Jarava, Ruiz and Pay., was founded on S. garava, Kunth
(S. eriostachys, Cav., S. papposa, Nees), a widely-spread West-
American species, to which the small spikelets in a long narrow
dense panicle, with the flowering glumes crowned under the awn

MR. G. BENTHAM ON GRAMINER. 81

by a pappus-like ring of long hairs, give a very peculiar aspect ;
but precisely similar flowering glumes are observable in several
South-American species with very various habits. In the Euro-
pean S§. pennata, Linn., and a few other American as -well as
Old- World species, the awn itself is (almost entirely, or for a short
distance above the base) plumose with long spreading hairs.
Lasiagrostis, Link (Achnatherum, Beauv.), was proposed as a
genus for the European S. Calamagrostis, Wablenb., and extended
by Nees and Trinius to several African and Asiatic species, only
differing from other small-flowered Stipe in the flowering glume
itself being plumose with spreading hairs, either below the middle
or in its whole length; and in S. mongholica, Trin., forming the
genus Ptilagrostis of Grisebach, these hairs extend to halfway up
theawn. S. verticillata, Nees, from Australia, and Apera arundi-
nacea, Hook. f., from New Zealand, two plants closely resembling
each other, though specifically distinct, connect Stipa with Mueh-
lenbergia. They have the inflorescence and small spikelets of the
latter genus; and in S. verticillata the awn is generally persistent,
though the articulation is distinctly traceable on the flowering
glume; in S. arundinacea the awn is very deciduous; in this
species there is usually but one stamen, whilst in 8. verticilata
there are the normalthree. S. rariflora (Muehlenbergia rarifiora,
Hook. f.), from Antarctic America, is another species closely
allied to the above two ; and all three appear to be better placed
under Stipa than under Muehlenbergia.

3. Oryzopsis, Mich. (Urachne, Trin.), is a genus of about four-
and-twenty species, from the temperate and subtropical regions
of the northern hemisphere or from extratropical South America,
very rare within the tropics, most of them often regarded as
awned species of Milium, but really more nearly connected with
Stipa, from which they chiefly differ in the broader fruiting glume,
often oblique at the top, the awn usually short, slender, and
twisted, and very deciduous. The genus divides readily into
three sections, regarded by some as distinct genera, but all united
into one by Trinius and others. 1. Piptatherwm, Beauv., com-
prises the Old-World species, often included in Miliwm as a
section, with awned glumes, and really connecting in some mea-
sure the two genera. The obliquity of the fruiting glume 1s
much less marked than in the typical species of Oryzopsis; and
the rhachilla of the spikelet is glabrous. 2. Hworyzopsis or
Oryzopsis proper, including the proposed genera Caryochloa,

82 MR. G@. BENTHAM ON GRAMINER.

Spreng., Piptochetium, Pres], and Nassella, E. Desv., is entirely
American, with the typical character of the genus, and the rha-
chilla bearing aring of hairs under the flowering glume. 38. Erio-
coma, Nutt. (Fendleria, Steud.), differs from Huoryzopsis only in
the long silky hairs clothing the fruiting glume.

4. Mit1um, Linn., was formerly extended to several unawned
Panices with only two empty glumes, but is now reduced to five
or six European or temperate Asiatic species, one of which is
also spread over North America, all removed from Panicacee as
having the empty glumes persistent below the articulation. They
differ from Oryzopsis chiefly in their obtuse absolutely unawned
flowering glume. 5. Acracune, Benth., is a single dwarf tufted
dioecious grass from the higher mountains of Peru and Colombia.
The female individual, with only one spikelet terminal on the pe-
duncle, is fully described and figured in the last part of Hooker’s
‘Teones.’ The male plant, if correctly matched, of which I am
by no means certain, has a loose almost simple panicle with pre-
cisely the glumes of the female, but enclosing stamens only. In
the few specimens seen the leaves are much longer than in the
numerous females from various localities, which makes me rather
doubt the specific identity of the two.

6. Muzurenperata, Schreb., has nearly sixty known species,
chiefly American, extending from the Andes of South America
over the northern continent generally, with a very few from
central or eastern Asia. They connect, in many respects, Stipa
with Agrostis. In general they come very near in technical
character to the smaller-flowered Stipe, differing in the still
smaller spikelets with thinner though still closely appressed and
narrow fruiting glumes, and usually with a more or less hairy
rhachilla. From Agrostis and its immediate allies they may be
readily distinguished by this narrow appressed fruiting glume
with a terminal never dorsal awn; a very few unawned species
are scarcely separable from Epicampes, except by the shape of the
glume. There is a considerable variety in the inflorescence and
in the proportions of the glumes, but nothing definite enough to
establish good sections, although several separate genera have
been proposed. In the original WZ. diffusa, Schreb., and its im-
mediate allies, the panicle is usually long, narrow, and dense,
and the lower empty glumes are very minute; whilst in Trinius’s
proposed section Aeroxis both the lower glumes or one only of
them are nearly as large as the flowering one; but throughout the

MR. G. BENTHAM ON GRAMINER. 83

genus the relative size of these glumes appears to vary almost
from species to species. Vaseya, Thurb., is a Californian species,
M. comata, Thurb., closely resembling the more common MW. syl-
vatica, Torr., except in the long hairs surrounding the thin
flowering glume. Podosemum, Desy. (Trichochloa, Beauv.), com-
prises a number of elegant species, in which the spreading panicle
has a number of small long-awned spikelets on long capillary
branches and pedicels. Tosagris, originally separated by Beau-
vois from Podosemum on account of the long hairs on the back
of the flowering glume, was subsequently reunited with it by the
author himself. Clomena, Beauv., is M. clomena, Trin. (AL nana,
Benth.), a dwarf Andine species, in which the second empty
glume is the largest of the spikelet, and rather broadly three-
toothed. The same character is observable in IL gracilis, Trin.,
forming Nuttall’s genus Calycodon.

7. BRACHYELYTRUM, Beauv., is a single North-American species,
very near to some species of Stipa; but the rhachilla is produced
beyond the flowering glume into a little bristle, sometimes bearing
a minute rudimentary glume, which does not occur in any other
species of the subtribe. 8. PERre1zema, Presl, contains three
or four tropical or subtropical American species, with much
of the habit and many of the characters of Muehlenbergia dif-
fusa, but with the empty glumes awned as well as the flowering
one.

Our second subtribe, PaLrorpEH, is chiefly characterized by the
inflorescence. The panicle is condensed into a globular or oblong
head or cylindrical spike; the rhachilla is, in a few species only,
produced beyond the flower into a small bristle; the flowering
glume either is awnless or bears one or three terminal awns, and
when in fruit is thinner than in Stipes, more loosely enclosing
the grain as in Euagrostee. The following seven genera, or most
of them, have already been placed in juxtaposition by various
Agrostologists.

9. Lycurus, H. B. K. (Pleopogon, Nutt.), consists of two closely
allied American species, perhaps varieties of a single one, readily
known by the empty glumes as well as the flowering one awned,
as in Perieilema, the lowest one having usually two or even three
awns. The long dense cylindrical spike (or spike-like panicle)
with sterile spikelets intermixed with the perfect ones brings the
genus in connexion with the subtribe Sesleries of Festucex ; but
there is never more than a single flower in the spikelet,

84. MRE. G. BENTHAM ON GRAMINER.

10. Ecutwopocon, Beauv. (Hystericina, Steud.), a single Aus-
tralian and New-Zealand species, has likewise sterile spikelets
intermixed with the perfect ones; but the empty glumes are
awnless, and the flowering one three-lobed with the middle lobe
produced into a long awn. 11. Dirtopogon, Br. (Dipogonia,
Beauv.), also a single Australian species, has a short awn to
the empty glumes and three to the flowering one, of which the
central one is long and twisted. 12. Ampurpogon, Br. (Agopo-
gon, Beauy., not of Willd., Pentacraspedon, Steud.), five Austra-
lian species, has the flowering glume deeply three-lobed and fre-
quently awned, and the palea also with two rigid almost awn-like
lobes. Gamelythrum, Nees, is the A. turbinatus, Br., separated
from Amphipogon only on account of a more distinct elongation
of the rhachilla between the outer glumes and the flowering
one.

18. Hetgeocutoa, Host (Pechea, Pourr.), contains seven or
eight Mediterranean species, of which one or two are widely
dispersed over Europe and Central Asia. Kunth referred them
to asection of Crypsis ; and Host himself subsequently assented
to the union, probably misled by an apparent resemblance of some
varieties of H. schanotdes to the true Crypsis aculeata; but the
resemblance is apparent only, the two genera are as essentially
different in inflorescence as in the structure of the spikelets.
The axis of inflorescence, or receptacle, in Crypsis is a flat disk ;
in Heleochloa it is a more or less elongated linear rhachis, cylin-
drical even in those varieties where the spike-like panicle is con-
tracted into a sessile head. In Crypsis the empty glumes are
above the articulation and fall off with the spikelet, and the
glumes are quite those of Oryzes without any two-nerved palea ;
in Heleochloa the empty glumes persist below the articulation,
and the glumes and palea are entirely those of Phleoidee; and
although in the commonest species the spikelike panicle or head
is short and sessile, yet there are others where it is long, narrow-
cylindrical, and pedunculate. Rhizocephalus, Boiss., founded on
Crypsis pygmea, Jaub. and Spach, makes, with C. ambigua and
C. crucianelloides of Balansa, a very good section of Heleochloa,
distinguished by the dwarf tufted habit and the spikelets almost
echinate with the rigid points of the glumes. Beauvois gave the
same name Heleochloa to a supposed genus, apparently made up
of a Sporobolus and a Phleum.

14. Mairtzs, Parlat.,is the Phalaris erypsoides, Dury., a dwarf

MR. G. BENTHAM ON GRAMINE. 85

tufted Greek plant, with the spikelets flat as in Phalaris, but
otherwise showing nearly the structure of Phlewm. 7

15. Parvum, Linn., about ten species from the temperate and
northern regions of the northern hemisphere or from Antarctic
America, is a well-known and already well-defined genus. It has
been proposed to separate generically Chilochloa, Beauv. (Achno-
don, Link), for the few species in which the rhachilla is produced
beyond the flower into a minute bristle; the character, however,
is in this instance very trifling and uncertain. Achnodonton,
Beauv., is P. tenue, Schrad., for which I can find no separate
generic character. The anomalous Phalaris trigyna, Host, appears
to have been an individual specimen of Phlewm Micheli, All.,
having abnormally three style-branches instead of two.

In a third small group or subtribe, SporoBoLEs, I should pro-
pose to place Sporobolus itself, with the three monotypic genera
Mibora, Coleanthus, and Phippsia. The subtribe is not very
clearly defined; but my previous endeavours to associate Sporo-
bolus with Milium and Isachne, to which I shall recur further on,
proved still less satisfactory. The plants now grouped together
have small paniculate or almost racemose spikelets, awnless glumes,
no continuation of the rhachilla beyond the flower, and the ripe
grain only half enclosed in and readily falling away from the
glume—characters sometimes well marked, but in some species
rather vague.

16. Mipora, Adans. (Chamagrostis, Borkh., Sturmia, Pers.,
Knappia, 8m.), is a dwarf slender tufted European annual, with
a simple spike and the lower empty glume at least as long as the
flowering one. 17. Connanruvs, Seid., isa minute annual, first
found in Bohemia, then in Norway, and more recently gathered
in the island of Sauvies at the mouth of the Oregon in North-
west America. It is very near Phippsia and Sporobolus; but the
lower empty glumes are entirely deficient. It was first disco-
vered by Seidel, and distributed by him under the name of Cole-
anthus subtilis; but Trattinick in publishing it (as reported by
Reemer and Schultes) retained only Seidel’s specific name, changing
the genus to Schmidtia. Roemer and Schultes in their ‘ Systema’
restored Seidel’s name, which Sternberg, rather later, changed
again both generically and specifically to Sehmidtia utriculosa.
Under these circumstances Seidel is now considered to have pub-
lished his Coleanthus subtilis sufficiently for general adoption,
more especially as another very different genus of Grasses has

86 MR. G. BENTHAM ON GRAMINER.

received the name of Sehmidtia, 18. Purppsta, Br., is a dwarf
paniculate slender Arctic grass, chiefly distinguishable from Spo-
robolus by the minute lower empty glumes.

19. Sporozsotus, Br. (Vilfa, Beauv., Agrosticula, Raddi, Tria-
chyrium, Hochst., Oryptostachys, Steud.), is now a genus of about
eighty species, spread over the warmer and temperate regions of
both the New and the Old World, mostly, however, American, with
a very few European or Asiatic. Included by the older authors
in Agrostis, it has since been universally acknowledged as dis-
tinct, though different characters have been assigned to it.
Beauvois, who attached primary importance to the presence or
absence of the awn, referred to it all the unawned species of the
old genus Agrostis. Brown, who first pointed out the differences
in the fruit, took as the principal character the loose membra-
nous pericarp readily detachable from the seed; but this, though
very conspicuous in S. indicus, Br., and in some other species, is
not apparent in the dried state in several others; and in &. virgi-
cus, Kunth, and others, it is only when soaked that the pericarp
can be detached. On this account it has been attempted to
establish two genera, Vilfa and Sporobolus ; but the character is
far too indefinite, as well as uncertain, to be available even for
sectional separation. As a whole, Sporobolus is chiefly distin-
guished from Agrostis by the total absence of any dorsal awn,
and by the grain so loosely enclosed in the glume that it usually
protrudes from it when ripe, and often falls away. The palea
also generally splits readily into two, and in some species is
even at the time of flowering divided to the base, a character
which Grisebach, who only observed it in an Argentine species,
was induced to take as that of a new genus Diachyrium ; but it
exists in many other species; and this divided palea has been
more than once described, and even figured (as in T. Nees’s
“Genera Flore Germanice ’), as a two-valved pericarp, a character
unknown in Gramines. Brown’s name Sporobolus was rejected
by Beauvois, Trinius, and others on the supposition that the genus
is identical with the older established Viifa, Adans. That, how-
ever, is a mistake. Adanson’s character of Vilfa is so vague
that it cannot be identified by that alone; but in his index he
fixes it by quoting two European species, which are certainly both
of them true species of Agrostis.

Two North-American species of Sporobolus, S. compressus,
Trin., and S. serotinus, A. Gr., are exceptional, not only in the

MR. G. BENTHAM ON GRAMINER. 87

genus, but in the whole tribe of Agrostex, by the spikelets con-
taining occasionally two flowers (without awus or continuation of
the rhachilla) as in Isachnew and in some species of Aira and Col-
podium ; but the small spikelets and carpological characters are
quite those of Sporobolus.

There remain for the Evacrostea, or fourth and last subtribe
of Agrostex, about sixteen genera, of which the general character
is a dorsal usually twisted awn on the flowering glume, the grain
neither so closely enveloped in the fruiting glume as in Stipes,
nor so readily exposed as in Sporoboles, and the spikelets usually
small, loosely paniculate, very rarely condensed into a head as in
Phleoidew ; but there are exceptions to every one of these charac-
ters, and the limits of the larger genera are so vague as to render
this portion of the genera of Graminee the least satisfactory of the
whole series. Of the sixteen following genera, the first seven
show uone of that continuation of the rhachilla beyond the flower
which in the others takes the form of a glabrous or hairy bristle
rarely reduced to a mere tubercle ; the last four of the series, as
well as Triplachne, have, besides the dorsal awn, two or four teeth
to the glume, sometimes produced into straight awns. A few spe-
cies or monotypic genera have no awn to the flowering glume,
but otherwise in the structure of the spikelet are nearer to Agrostis
than to Sporobolus.

20. Errcampns, Presl, about sixteen species from Mexico and
the South-American Andes, probably reducible to about two
thirds of that number, is a genus most embarassing to the syste-
matist; for it seems to connect Muehlenbergia and Sporobolus with
Agrostis. The chief general feature is the long narrow dense
panicle with very numerous rather small spikelets, the awn
of the flowering glume, when it exists, much smaller, than in
Muehlenbergia and often not quite terminal; the unawned species
distinguished from Sporobolus by the fruiting glume and grain
nearly those of Agrostis, and from the latter genus by the inflo-
rescence and by the awn when present being very small and
almost terminal. Several of the published species, however, are
unknown to me; and a further study may require considerable
modification of the generic character and limits. Crypsinna,
Fourn., which appears inseparable from Epicampes, is founded
on the BE. macroura (Cinna macroura and C. stricta, Kunth), a
Mexican species remarkable for the very long, narrow, almost
spikelike panicle. Cinna macroura, Thurb., from California, is a

88 MR. G. BENTHAM ON GRAMINEE.

distinct species (Z. rigens, Benth.) with a still longer and nar-
rower rigid spikelike panicle often interrupted; and the spikelets
are small, with membranous glumes, as in the typical 2. stricta,
Pres], but awnless or nearly so.

21. Bavcuea, Fourn.,a single Mexican species unknown to me,
might perhaps be reduced to Epicampes, from which it is said to
differ chiefly in a great inequality of the two empty glumes.

22. Agrostis, Linn., even after being shorn of a number of
heterogeneous plants ascribed to it at various times from general
aspect, and after the suppression of numerous names given to
local representatives of cosmopolitan species, is still a genus
of nearly a hundred species, generally spread over nearly the
whole world, but especially common in the temperate regions
of the northern hemisphere. Among the Euagrostee without
any continuation of the rhachilla they are generally known
by the absence of those peculiarities which have induced the
separation of several of the preceding as well as of the follow-
ing genera, by the thin short broad flowering glume with the
dorsal awn below the middle, and by the palea not more than
half as long as the glume, and often quite minute or even defi-
cient; the panicle is also usually loose, very elegant, with nume-
rous small spikelets or almost capillary branches and pedicels.
There are, however, here, as elsewhere, exceptional species which
defy all neat classification ; even the dorsal awn is sometimes re-
duced to a minute tubercle, or only to be guessed at by the abrupt
termination of the central nerve of the glume. The American
species in which the palea is minute or deficient formed Michaux’s
genus Trichodium, which has been extended to the European
A. canina, Linn., and others with that character, to which Beau-
vois restricted the Linnean name Agrostis, whilst he gave the name
of Agraulus to the species in which the palea is more developed.

23. Cuzmtunvus, Link, is a single Spanish annual, much like
some species of Agrostis, but anomalous in the group in having
the lowest (empty) glume larger than the others, and produced
into a long awn, whilst the flowering one, though shaped as in
Agrostis, has no awn. The inflorescence is also peculiar, each
branch of the panicle terminating in the three shortly pedicellate
spikelets.

24, Arcragrosris, Griseb., is a single arctic species, referred
by Brown doubtfully to Trinius’s genus Colpodiwm, but which
appears to be more nearly allied to Deyeuwia. It is, however,

MR. G, BENTHAM ON GRAMINES. 89

one of those plants which, by irregularity in some characters
usually very important, is very difficult to place satisfactorily.
The habit and size of the spikelets are more those of Poa than of
Agrostis ; but, in the great majority of specimens, the one-flowered
spikelets without any continuation of the rhachilla are quite
those of Agrostew, and the palea is fully the length of the glume
asin Deyeuxia. Very rarely specimens have presented themselves
with a minute continuation of the rhachilla ; and Brown, in a single
Melville-Island specimen, found it to bear an empty glume or
second flower, thus showing a connexion, and possibly, in the spe-
cimen mentioned by Brown, a hybrid between Arctagrostis and
Poa alpina.

25. Catamagrostis, Adans., as now limited, comprises four or
five species from Europe and northern and central Asia, of which
one has also been found in South Africa, possibly, but not cer-
tainly, introduced there. Some authors extend the genus so as
to include the greater part of Deyeuxia, and indeed all the
Euagrostex with a hairy rhachilla; but it seems more natural if
confined to the typical species, which, like Agrostis, have no con-
tinuation of the rhachilla or rarely a very slight one, and bear on
the flowering glume a fine dorsal awn, rarely reduced to a minute
point. They differ from Agrostis in the ring of long hairs sur-
rounding the flowering glume, and generally in their tall almost
reed-like habit, whence their generic name, and on which account
they have often been placed in juxtaposition with Arundo. They
appear, however, to be in every respect true Agrostex ; and there
are two species, C. tenella, Kunth, and C. olympica, Boiss., which
are almost intermediate between Calamagrostis and Agrostis, espe-
cially as a few species of true Agrostis are not entirely without
hairs on the rhachilla.

26. Cryna, Linn. (Abola, Adans., Blattia, Fries), is limited by
modern authors to two species from the northern regions of
Europe and America, with the tall reedlike habit of the larger
species of Calamagrostis, but with a glabrous rhachilla, and re-
markable in the tribe by the palea having only one nerve, although
there is every reason to believe that it is a true palea, the appa-
rently single nerve being due to the consolidation of two. Both
species appear also constantly to have but one stamen in the
flower. Some botanis!s unite the two; but from dried specimens
they appear quite distinct. Amongst other minor points, the
original C. arundinacea, Linn., has generally a minute continua-

90 MR. G@. BENTHAM ON GRAMINES.

tion of the rhachilla, which I have never found in C. pendula,
Trin. (C. expansa, Link, C. latifolia, Griseb.). Several American
or other grasses published as species of Cinna are now referred to
Epicampes or Deyeuxia. é

27. GasTRipium, Beauv., has two species from the Mediterra-
nean region, one of them also found in tropical Africa and in
extratropical South America, but possibly introduced only in the
latter locality. They have the small spikelets of Agrostis, but a
narrower closer panicle, and are remarkable for the outer glumes
rather hardened, shining, and ventricose at the base, whence the
generic name. The older authors included them in Mlium on
account of that hardness in the glumes.

28. Cuztorropis, Kunth, is a single Chilian species which
some would unite with Agrostis, and might well have been joined
to Epicampes, but that the rhachilla is produced beyond the
flower into a rather long hairlike seta.

29. Triptacune, Link, is the Gastridium nitens, Coss. et Dur.,
a single Mediterranean species, with the habit, but not the ven-
tricose glumes, of that genus, and differing both from that and
from Agrostis in the flowering glume bearing a short awn-like
point on each side of the awn.

30. Aprra, Adans. (Anemagrostis, Trin.), has two very closely
allied European species extending into Western Asia, with the
technical character very nearly of Deyeuxia, but with the elegant
panicle and numerous small glabrous spikelets of many species of
Agrostis, in which they are still included by some under the name
of Agrostis spica-venti. The New-Zealand plant described by
Hook. f. as an Apera is now transferred to Muehlenbergia.

31. Cryvagrostis, Griseb., from Tucuman in South America,
is unknown to me, but is said to differ from Deyeuata in having
the spikelets unisexual by abortion. It should most probably be
incorporated in that genus.

82. Duyuuxta, Clarion (Lachnagrostis, Trin.), has now nearly
a hundred and twenty species, dispersed over the temperate or
mountain regions of the globe, particularly numerous in the
Andes of South America, and extending northwards to the Arctic
circle and southwards to the extreme end of South America. It
is in some respects polymorphous, running on the one hand
almost into Agrostis, to which some species have been referred,
and on the other into Calamagrostis, with which the northern
species have been often united. It differs from both in the pro-

MR. G@. BENTHAM ON GRAMINED. 91

longation of the rhachilla into a bristle or stipes usually, but not
always, hairy, and from Agrostis in the larger spikelets, with the
palea nearly as long as the glume, and the usually hairy rhachilla.
There are, however, a few species where one or another of these
characters fail; and one or two scarcely differ from Apera except
in the habit and in the awn being more decidedly dorsal. Bro-
midium, Nees (to which belong Didymocheta, Steud., and Chame-
calamus, Meyer), contains a few Andine Chilian or Australian
species, which, with the minute glabrous and perhaps not con-
stant continuation of the rhachilla, might almost as well be
transferred to Agrostis, but that they have rather the habit of
Deyeuxia. Relchela, Steud. (Agrostis sesquivalvis, E. Desy.),
Cinnastrum, Fourn. (at least as to Deyeuxia poeformis, Kunth),
Deyeuxia mutica, Wedd., and D. breviglunis, Benth., with a few
other South-American species, form a little group with a glabrous
rhachilla and the awn reduced to a small point. In Acheta,
Fourn., two Mexican species, the awn appears to be deficient ;
but all the other characters are those of the typical Deyeuxie with
a hairy rhachilla, to which I would also refer the Agrostis equi-
valvis, Trin., forming Grisebach’s section Podagrostis.

33. Ammopuita, Host (Psamma, Beauv,), comprises four spe-
cies, two of them widely spread over the northern hemisphere
chiefly near the sea, and two confined to North America. They
are distinguished from Deyewaia by their tall habit, their usually
dense inflorescence, and especially by their larger paleaceous
glumes.

34. Dicuetacuye, Endl., two Australian or New-Zealand spe-
cies, with a narrow dense panicle, differs from Agrostis and its
allies in the flowering glume scarcely smaller than the outer
empty ones and often toothed, and in the long dorsal awns giving
the inflorescence a fine bristly aspect. J

85. Triserarra, Forsk. (Anomalotis, Steud.), is a maritime
Syrian and Egyptian plant, very near Dichelachne, but still more
bristly, the lateral teeth of the flowering glumes being produced
into fine straight awns, whilst the dorsal one is longer and flexuose.
Labillardiére and Delile both mention two fertile flowers in the
spikelet. I bave only been able to find one in several specimens
examined, all from Alexandria; possibly they may have consi-
dered the rather long continuation of the rhachilla as a second
flower.

36. Penrarogon, Br., is a single Australian species, with four

92 MR. G. BENTHAM ON GRAMINED.

straight awns to the flowering glume, besides the long rigid
twisted dorsal one, which, as well as the single flower, removes
the genus from the Pappophoree.

37. Lacurvus, Linn., is a well-known widely spread Mediter-
ranean grass, which, like Zrisetaria, has two slender awns to
the flowering glume besides the more rigid dorsal one, but is
well marked by the capitate inflorescence, to which the long hairs
of the linear plumose empty gluines give a peculiar soft silky
aspect.

Tribe IX. Isacunea.

This small tribe is a modification of the subtribe I proposed
in the ‘ Flora Australiensis ’ under the name of Milies, and which
I distinguished from Agrostez by the absence of the dorsal awn,
and from Festucee by the single or two equal flowers in each
spikelet ; and I included in the group both Miliwm and Sporo-
bolus. Since I have worked up the. Agrostee of the northern
hemisphere, however, I find that the presence or absence of the
dorsal awn is much more uncertain than I had thought, that
Milium cannot be removed far from Oryzopsis, and that Sporo-
bolus must be referred back to Agrostee. But there remain a
group of genera, nearly related both to Agrostes and to Avene,
but never showing the dorsal awn so general in those tribes,
and enclosing in each spikelet two equal flowering glumes and
flowers, apparently inserted at the same point without any deve-
lopment of the rhachilla between them (except in Calachne) and
never avy continuation beyond the flowers. The two flowers are
both hermaphrodite and fertile ; or occasionally only one of them,
usually the upper one, is female or sterile. The tribe thus limited
would consist of the following seven genera :—

1. Prronacnne, Nees, subsequently republished by the same
author under the name of Chondrolena, is a South-African annual
with an almost simple terminal spike, distinguished by the outer
empty glumes as long as the flowering ones, with a rigid pec-
tinately-toothed cartilaginous keel. Ktenosachne, Steud., is most
probably the same plant.

2. Isacune, Br., comprises about twenty tropical or subtro-
pical species, chiefly from the Old World, but including a few
American ones. The small spikelets with the loosely paniculate
inflorescence and more or less hardened fruiting glumes give them
the appearance nearly of some species of Panicum, to which

MR. G. BENTHAM ON GRAMINESA. 93

genus some species of Isachne have been referred, but from which
they constantly differ in the empty glumes persistent below the
articulation, andin the two flowers both hermaphrodite or female,
though one may be occasionally sterile. Graya, Nees (judging
from the reference to Wight, but not from Steudei’s characters)
is Isachne pulchella, Roth (Panicum bellum, Steud., P. malaccense,
Trin.). Panicum Gardneri, Thw., which, as the author observes,
is closely allied to Zsachne Walkeri, Wight, appears to me to be
strictly congener with that species, although one of the flowers of
the spikelet is frequently, but not always, sterile. It seems to be
the same as Isachne nilaghiriea, Mochst.

3. Zenxeria, Trin., very well described and figured in the
‘Linnea,’ vol. xi., now contains two species from the East-Indian
peninsula and Ceylon, both very near Isachne, but with membra~
nous fruiting glumes. -Amphidonax Heynei and A. tenella, Nees,
do not differ from the typical Z. elegans, Trin. The second spe-
cies is Z. obtusiflora (Amphidonax obtusifiora, Thw.). The ori-
ginal genus Amphidonax of Nees was founded on a species of
Arundo.

4. Micnarra, F. Muell., is a single North-east Australian spe-
cies recently figured in Hooker’s Icones. 5. Caxnacuye, Br.,
comprises three Hast-Indian, Chinese, or East-Australian species—
C. pulchella, Br., C. perpusilla, Thw., and C. simpliciuscula, Munro
(Isachne simpliciuscula, Wight et Arn.), which, as above observed,
are anomalous in the tribe by a slight extension of the rhachilla
between the flowering glumes. 6. Arropsis, Desy., restricted to
the single West-Mediterranean A. globosa, is a pretty little
annual, formerly placed in Miliwm on account of the hardening
of the glumes, or in Aéra, which it resembles in many respects.
Tt shows, however, all the characters of Isachnezx, and is indeed
technically nearly allied to Isachne itself ; but the two semiglobose
fruiting glumes, closely appressed to each other by their flat
faces, give the spikelets the peculiar globular shape expressed by
the specific name.

7, Ertacune, Br., comprises twenty-two species, two of them
endemic in tropical Asia, the remainder Australian, of which one is
also in East India. They differ from Jsachne generally in their
rather larger spikelets, and especially in the long hairs on the back
or margins of the flowering glumes, and sometimes in the fine
straight awns terminating the flowering glumes, or even the teeth
ofthe pales. Megalachne, Thw. (not of Steud ), is Briachne triseta,

LINN. JOURN.— BOTANY, VOL. XIX. I

94 MR. G. BENTHAM ON GRAMINEA}.

Nees, in which these awns are particularly conspicuous. The
African species referred by Nees to a section Achneria of Eriachne,
form a distinct genus of Avene, for which Munro has retained
this name Achneria, the original genus Achneria of Beauvois having
been proposed for those true Australian species of Eriachne which
have no awn or only a very small one.

Tribe X. AVENE.

This tribe has been more generally recognized and subjected to
less variations than most of the others. Its general characters—
the paniculate inflorescence, the spikelets with two or more per-
fect flowers, the rhachilla”’produced beyond the upper flower, and
a twisted awn to the flowering glume either dorsal or terminal
between the two lobes or teeth of the glume—suffer fewer excep-
tions than usual. Aira alone has no continuation of the rhachilla ;
and Anisopogon alone has only one perfect flower in the spikelets.
Of the following sixteen genera, the first eleven have the awn
dorsal and the lowest flower hermaphrodite ; the next three have
the male or sterile flower below the perfect one; and the last two
have the lowest flower hermaphrodite and the awn terminal.

1. Arra, Linn., was once made to include Corynephorus, Des-
champsia, and indeed almost all the Avene with loosely panicu-
late inflorescence and small two-flowered spikelets, but has since
been so thoroughly dismembered by various European botanists
as not to leave a single species to represent the old Linnean
name. Taking, however, the widely spread A. caryophyllea, Linn.,
as a genuine type, and adding to it five or six European species,
we have a natural genus of elegant, slender, mostly annual grasses
with fine filiform leaves, the small spikelets always two-flowered
without any continuation of the rhachilla beyond the upper flower,
the dorsal awn of the flowering glumes rarely wanting, and the
ripe grain often adhering to the palea; the latter character, how-
ever, is always uncertain. These six or seven species have all
been made the types of supposed distinct genera. A. caryophyllea,
Linn., and A. precox, Linn., considered as typical Aira by Par-
latore and others, form the genus Fwssia, Schur, in which the two
flowers are closely contiguous and the flowering glumes usually
awned. Fiorinia, Parlat., is the A. Tenorii, Guss., distinguished
by the absence of the awn; but Gussone has shown that it varies
with or without the awn. dAntinoria, Parlat., is the A. agrostidea,
Lois., with the rhachilla more or less lengthened between the

MR. G. BENTHAM ON GRAMINEA. 95

flowers, and the glume usually unawned ; but A. pulchella, Willk.,
with the glume awned, cannot be otherwise distinguished from
it. Periballia, Trin., is the A. involucrata, Cay., in which the
two flowers are as in A, agrostidea rather distant from each other,
the lowest flowering glume unawned, the upper one awned, but
both flowers hermaphrodite, as in the rest of the genus. The
inflorescence of this species is rather peculiar; the lowest whorl of
branches of the panicle are usually without any or with only very
few spikelets, and were regarded by Cavanilles as an involucre:
but that is not always the case; I have seen some specimens
with spikelets on all the branches. A. sabulonum, Labill., from
New Caledonia, is a very doubtful plant. Labillardiére’s figure
is a good representation of the Australiasan form of Sporobolus
virginicus, except that the spikelets are drawn as two-flowered.
The specimens sent for Labillardiére’s plant by Pancher and by
Vieillard have only one flower in the spikelet.

2. CoryyrpHorus, Beauv. (Weingartneria, Bernh.), comprises
two European grasses, extending into North Africa and more
sparingly to the Levant, with the continuation of the rhachilla
of Deschampsia, but readily distinguished by the peculiar articu-
late club-shaped awn of the flowering glumes.

8. Drscuampsia (Campelia, Link) is a genus of about twenty
species, from the temperate or colder regions of both the New
and the Old World, sparingly represented in mountain regions
within the tropics. It bears the same relation to Aira that
Deyeuxia does to Agrostis; the plants are usually perennial and
stouter than in Aira, the spikelets larger, and the rhachilla is
produced beyond the upper flower into a bristle often bearing a
tuft of hairs, and sometimes an empty glume on even a male
flower ; the flowering glumes are also frequently more or less
denticulate. No less than six of the species have been proposed
as distinct genera :--Vahlodea, Fries, is D. atropurpurea (Aira
atropurpurea, Wahlenb.), a northern species, with the flowering
glume not at all or only very minutely denticulate, otherwise
quite a Deschampsia. Avenella, Parlat. (Lerchenfeldia, Schur),
is the common D. flexuosa (Aira flexuosa, Linn.), with the flower-
ing glume surrounded by hairs. The grain is said by Parlatore
to adhere to the palea, which may be sometimes, but is certainly
not always, the case. Monandraira, Em. Desy., includes two
Chilian species, Zrisetwm Berteroanum and Z. aireforme of Steudel,
separated from Deschampsia as having but one stamen to the

12

96 MR. G. BENTHAM ON GRAMINED.

flower; but in D. Berteroanum I have sometimes found two
stamens, and in the evidently nearly allied D. antarctica, Hook. f.,
the stamens, though usually three, are sometimes two only.
Airidium, Steud., is a species from the Straits of Magellan which
Iam unable to distinguish from the D. antarctica. tytidosper-
mum, Steud., is founded on specimens of a Deschampsia closely
allied to, if not identical with, the common JD. cespitosa,in which
a erub has taken possession of every spikelet remaining in the
panicle, and has been mistaken by Steudel for the caryopsis, and
actually described as such. Peyritschia, Fourn., is D. kelerioides
(Aira hoelerioides, Peyr.), which I have not seen, but which, from
Peyritsch’s elaborate description, must be very near to D. ant-
arctica, Hook. f., D. nitida, Presl, and D. holciformis, Presl.

4, AcunerIaA, Munro, contains eight South-African species,
with one from south-eastern tropical Africa, referred by Nees to
Eriachne, and by Kunth to Airopsis, but evidently more nearly
related to Deschampsia.

5. Moracuyroy, Parlat., is a single species from the Cape-
Verd Islands, which we have not at Kew, and of which I have
therefore been unable to verify the character given by Parlatore.
The specimen he described most probably remained in Webb's
herbarium, now deposited at Florence.

6. Hotcvs, Linn., formerly included two very different groups
of grasses ; and Brown specially retained Linnus’s name for that
one which now forms the genus Sorghum in Andropogonex,
whilst all modern botanists restrict the genus Holcus to the other
group, consisting of about eight European or African species,
chiefly western, of which one or two are common weeds in various
parts of the world. All are nearly allied to Deschampsia, but
have the upper flower of each spikelet male with an awned glume,
and the lower one unawned and hermaphrodite. Two Spanish
species have been added by Boissier, H. grandiflorus and H. cas-
pitosus; but as they have both the flowers hermaphrodite and
awned (whence the sectional name Zomalachne), they should
rather be transferred to Deschampsia, although they may have
the peculiar soft habit of the common species of Holcus.

7. Trisrrum, Pers., is now known to comprise nearly fifty
species, ranging over the temperate or mountain regions of both
the New and the Old World. All are very near to the section
Avenastrum of Avena, but differ generally in the flowering glume
decidedly toothed at the apex, the two teeth often produced into

MR. G. BENTHAM ON GRAMINER. 97

straight awns, one on each side of the dorsal twisted one, and in
the grain glabrous or slightly pubescent at the apex without the
longitudinal furrow of Avena. The inflorescence is also usually
more dense than in that genus, with smaller, often shining spike-
lets. A few African or South-American species, however, such
as T. hirtum, Nees, and ZT. antarcticwm, Nees (which includes
Bromus antareticus, Hook. f., and Bromus bicuspis, Nees), closely
connect the two genera: the flowering glume is more rigid and
less keeled than in the true Zriseta, and the ovary is pubescent
at the top ;but the grain hag not the furrow of Avena. Tricheta,
Beauv., is the Trisetum ovatum, Pers., a species allied to 7. sub-
spicatum; but the spikelike panicle is more dense and ovoid, or
almost globular. Acrospelion or Acropselion, Bess., is Trisetum
distichophyllum, Beauv., not the Ventenata, Link, to which it is
referred in Lindley’s ‘ Vegetable Kingdom.’ ostraria, Trin.,
was made up of Trisetum neglectum, Roem. et Schult., and Keleria
phleoides, Pers.

8. Ventenats, Keel., has two species, V. avenacea, Keel., and
V. macra, Balansa, from the Mediterranean region and Central
Europe, differing slightly from Trisetwm in the longer, more rigid,
many-nerved glumes, and the absence of any dorsal awn on the
lower flowering one.

9. Avena, Linn., as limited by recent authors, comprises about
forty species, mostly from the temperate regions of the Old
World, with a few from extratropical North and South America,
and one or two of the annual ones cornfield weeds in other coun-
tries. It is generally characterized by the flowering glumes
rounded on the back and several-nerved, with a dorsal twisted or
bent awn, and by the ripe grain furrowed in front and more or
less adhering to the palea, butis divisible into two sections almost
marked enough in habit as well as character to be raised to the
rank of genera. In 1. Crithe, Griseb., the species are all annual,
usually tall, with a loose panicle of large pendulous spikelets, each
containing no more than two fertile flowers, and often only a
single one, and the lower empty glumes 7- or 9-nerved. This
section includes the common Oat, which has lost its dorsal ayn
probably as a consequence of long cultivation; for the plant is
unknown in a wild state, except here and there as an escape
from cultivation. In 2. Avenastrum, Koch (Helicotrichum, Bess.),
the plant is perennial, the panicle usually narrow, with erect or
rarely spreading spikelets with more than two perfect flowers,

98 MR. G, BENTHAM ON GRAMINES,

and the lower empty glumes with only one or three, or the second
rarely with five nerves.

10. Gaupry1a, Beauy., two species, has the spikelets of dvena
(Avenastrum) ; but they are singly sessile in the notches of the
articulate rhachis of a single spike, thus showing the inflorescence
of the tribe Hordes, to which Parlatore would remove the genus ;
but the dorsal twisted awn places it much nearer to Avena, from
which some authors would not generically separate it. The com-
mon G. fragilis, Beauv., is widely dispersed over the Mediter-
ranean region. The second species, G. geminiflora, J. Gay, was
proposed asa genus Arthrostachya, Link, from garden specimens of
unknown origin; it has since been detected by Seubert in the
Azores.

11. Aupurpromus, Nees, is a single Australian species, with
many-flowered spikelets. The grain is furrowed as in Avena, but
glabrous and free from the palea as in Zrisetum.

12. AnrHENATUERUM, Beauv., contains three European, North-
African, or Oriental species, often included in Avena, but differing
from that genus as well as from most Poacee in having, as in the
two following genera, the lower flower male and the upper one
fertile, though the rhachilla is produced beyond it as in other
Avenes.

18. Tristacnya, Nees (Jfonopogon, Presl), has eight species,
of which two are tropical American, the remainder African, tro-
pical or southern, one extending to the Levant. With the lower
flower male, as in Arrhenatherwm, they are readily distinguished
by the spikelets alwavs three together, sessile or equally pedi-
cellate at the ends of the branches of the panicle, and by the
long twisted awn of the flowering glume being terminal between
two lobes or straight awns. Amongst Nees’s African species,
T. simplex must be transferred to TLrichopterys.

14, Tricnorreryx, Nees (Loudetia, Hochst.),about ten African
species, of which one is also in Brazil, has the spikelets of Tri-
stachya; but they are scattered along the branches of the panicle,
not in terminal triplets. The only Brazilian species, not un-
common also in tropical Africa, 7. jlammea, has, as already men-
tioned, been rather negligently published and figured as an
Arundinella, of which it has none of the characters and not much
of the habit.

15. Antsorocon, Br., is a single West-Australian species,
differing from Danthonia in the large spikelets containing only a

MRE, G. BENTHAM ON GRAMINED. 99

single perfect flower. Nees added a second species from South
Africa which I have not seen; but from his description it can
scarcely be a congener. Kunth has figured three lodicules in
the Australian plant ; I have always found only two long lanceo-
late ones.

16. Danrnonta, DC., is now a polymorphous, almost cosmo-
politan, genus of nearly a hundred species, of which the greater
number, however, are South-African, all characterized by the
spikelets containing three or more perfect flowers, and by the
awn of the flowering glumes more or less twisted or bent and
usually flattened at the base, but terminal between two or four
teeth or straight awns. Notwithstanding considerable diversities
in habit, inflorescence, and in the size and teeth of the glumes,
no good natural sections have yet been proposed. Nees’s Hi-
mantochete (Streblochete, Hochst.), with the lateral lobes or teeth
of the flowering glumes entire and acute or awned, and Penta-
schiste, with the lateral teeth bifid and one or both teeth awned,
are purely artificial, and relate to the African species, all the non-
African ones being included in Himantochete. DeCandolle ori-
ginally proposed the genus for two European species, D. procum-
bens and D. provincialis ; Brown showed, however, that they could
not well be regarded as congeners, and removed the former to
his new genus Triodia. The D. provincialis therefore becomes
the type of the present large genus Danthonia, though it may be
somewhat anomalous when compared with the majority of the
African and Australian ones. DeCandolle’s chief character con-
necting his original species was the great length of the outer
empty glumes compared with the rest of the spikelet ; and this is
a general, though not quite a universal, feature of the enlarged
genus. Since Brown’s time the following genera have been pro-
posed, chiefly upon single species, with characters which appear
to be of little more than specific value :—Pentameris, Beauv., is
D. Thouarsii, Nees, from South Africa, with nearly the habit and
inflorescence of D. pallescens, Nees, but remarkable for the short
thick grain truncate at the top. Triraphis, Nees (not of R. Br.),
is D. radicans, Steud., from South Africa, nearly allied to D. erspa,
Nees. Chetobromus, Nees, contains a few South-African species,
in which one, or sometimes two, of the flowers in the spikelet are
imperfect. Monacather, Steud., is D. bipartita, F. Muell., an
Australian species, with the fruiting glumes hardened and oblique
at the base and bearing a ring of hairs under the lobes. Plin-

100 MR, GC. BENTIUAM ON GRAMINEZ.

tanthus, Steud., is founded on two Australian species, most pro-
bably of Danthonia, but which, from the evidently incorrect
character given, it is impossible to identify without seeing the
specimens. Crinipes, Hochst., is the Abyssinian species pub-
lished by A. Richard as D. abyssinica, Hochst., in which the outer
empty glumes are exceptionally shorter than the spikelet.

Tribe XI. CHLorIDEm.

This tribe is characterized amongst Poacexe almost exclusively
by the inflorescence. The spikelets are sessile in two rows in
unilateral spikes, the rhachis of which is neither articulate nor
notched as in Hordeew; and the spikes, sometimes solitary and
terminal, are more frequently several, either digitate at the end
of, or scattered along, the peduncle or axis of the single panicle.
The inflorescence is thus nearly that of Paspalum, whilst the
spikelets are those of Festucex, with the lowest or single perfect
flower hermaphrodite, and the awns, when present, terminal and
straight, not dorsal or twisted as in Agrostesw and Avenex. The
following twenty-seven genera are mostly, but not quite all, tro-
pical or subtropical; the first fifteen have one fertile and only
rarely a second male flower in each spikelet; the next ten have
two or more fertile flowers; all, except a few very small genera or
exceptional species, have the rhachilla continued beyond the
flowers, and often bearing one or more empty glumes. The last
three genera enumerated under the tribe are anomalous dicecious
grasses, connecting Chlorides with the subtribe Sesleriex of Fes.
tucex, but showing the inflorescence of the present tribe at least
in the male individuals.

1. Mrcrocuxoa, Br., comprises three species, of which two are
endemic in Africa and the third widely spread over the warmer
regions of the New as wellas the Old World. ‘They are slender
tufted grasses with filiform leaves and single slender terminal
spikes and small awnless one-flowered spikelets without any
continuation of the rhachilla.

2. Scua@nrreLpia, Kunth, is a single tropical-African species
with one to four erect spikes at the top of the peduncle; the
spikelets are one-flowered without any continuation of the rha-
chilla as in Microchloa, but not so small; and the flowering
glumes bearing long capillary awns, give the spikes an elegant
crinite aspect.

3. Cynovon, Pers. (Dactilon, Vill., Capriola, Adans., Fibichia,

MR. G. BENTHAM ON GRAMINE. 101

Keel.), a small but mixed genus, of which the typical species is
a common weed in most warm or temperate parts of the civilized
world. It has the slender spikes and small spikelets of Merochloa ;
but the spikes are several digitate at the end of the panicle, and
the yhachilla is produced beyond it into a small point or bristle.
Three Australian species have, however, been added to it with the
spikelets of Microchloa but with the inflorescence of Cynodon,
thus closely connecting the two genera. Persoon’s generic name
is far from being the oldest, but has been so long and so uni-
versally adopted, that the substitution of either of the others
for it would only breed confusion without the slightest advantage.

4, Harpecuioa, Kunth, has also two South-African species.
The spike is single, terminal, dense, and unilateral, often faleate ;
and there are usually one or two male flowers above the fertile
one, the glumes all unawned.

Of 5. Crznrum, Panz. (Monocera, Ell.), we have seven species,
of which four from North or South America, three from Africa
or the Mascarene islands. The spikes are solitary or rarely two
or even three at the end of the peduncle; the spikelets, though
elegantly pectinate as in Harpechioa, have a very different struc-
ture: the second empty glume is larger than the others, and bears
on the back a fine horizontal point sometimes reduced to a small
tubercle ; the third and fourth glumes are small and empty, or
only enclose a palea; the fifth or flowering glume ends in a fine
awn, and above it are one or two empty ones.

6. Enteropocon, Nees, was founded on an East-Indian grass
with asingle long, often incurved terminal spike ; otherwise very
near Chloris except in some minor points. It now includes
Ctenium Seychellarwm, Baker, from the Mauritius, which is scarcely
specifically distinct from the common East-Indian one, H. macro-
stachya, Munro (Chloris macrostachya, Hochst.), from Abyssinia,
and an unpublished species from Mayotte in Madagascar, Boivin
n. 8019, which may be thus characterized :—-Z. leptophylia, Benth.,
foliis augustissimis siccitate convoluto-subulatis, gluma florentis
integre arista gluma ipsa longiore. The habit and the long
unilateral spike are precisely those of the common Indian H. me-
licoides ; but the leaves are very much narrower and scarcely
flattened in the lower part, the spikelets rather larger, the flower-
ing glume nearly 3 lines long, and the awn about 3 inch, and,
at least in the spikelets examined, the flowering glume is quite
entire, scarcely free at the point from the awn.

102 MR. G, BENTHAM ON GRAMINER.

7. Cutoris, Sw., contains about forty species, dispersed over
the warmer regions both of the New and the Old World. It is
for the most part a natural genus, with two or more spikes digitate
at the end of the peduncle, the one-flowered spikelets in two
regular close rows as in the allied genera, the flowering glume
usually awned, and one, two, or more empty glumes above it; but
these characters are not constant, and the structure of the spike-
lets is somewhat polymorphous. C. monostachya, Pourr., from the
Mascarene islands, and C. wnispicea, F. Muell., from Australia,
are slender plants with only one or rarely two spikes, and the
flowering glume as well as the upper empty one are narrow and
awned. In C. aciculare, Br., and C. Roxburghiana, Edgew., from
East India and Australia, and C. radiata, Sw., from America
and Africa, the glumes are likewise narrow and awned, or the
upper empty one reduced to a mere awn, but the spikes are nor-
mally several and digitate. C. foliosa, Willd., has also a narrow
awned flowering glume; but the upper empty one is a double awn
(or two awnlike glumes), and the spikes are not so closely clus-
tered at the end of the peduncle, on which account Doell has
transferred the species to Gymnopogon, from which it appears to
me to be much further removed. In C. pumilio, Br., C. pectinata,
Benth., and C. divaricata, Br., all from Australia, the flowering
glume is distinctly two-lobed with the awn between the lobes.
In a considerable number of species the upper empty glumes are
broad and truncate at the top-—these empty glumes being several
in each spikelet in the Asiatic and Australian species, but one
only in the American ones. In all the preceding species the
flowering and upper glumes are awned; in five or six American
or African species forming the proposed genus Hustachys, Desv,
(Schultesia, Spreng.), both the flowering and the upper empty
glume are obtuse and truncate, but without any awn, or only a
minute point. They are, however, closely connected with the
typical species of Chloris through C. submutica, Kunth. C. villosa,
Pers., and C. macrantha, Jaub. et Spach, both of them described
in detail and figured by Jaubert and Spach, must be transferred to
Tetrapogon,as having their spikelets with at least two fertile flowers.

8. Tricuiorts, Fourn., comprises two Mexican and two extra-
tropical South-American species. They resemble Trisetaria in
their dense oblong crinite panicle and their three-awned flowering
glumes; but the panicle is composed of simple crowded or verti-
cillate spikes, and the spikelets, sessile in two rows on the rhachis

MR. G. BENTHAM ON GRAMINES. 103

with one to three empty awned glumes above the flowering one,
are quite those of Chloris. The two southern species had long
been indicated and named in herbaria as constituting an inde-
pendent genus (the one by J. Gay, the other by Munro); but
never having been published, we must adopt Fournier’s generic
name for the whole. The two southern species (from Chili,
Tucuman, and Buenos Ayres) are indeed so very near FZ. fasci-
culata, Fourn., that it will require close investigation to estab-
lish their specific differences. Another different-looking plant
from Tucuman (Tweedie), much smaller, with a loose inflores-
cence and short-awned spikelets, shows also the essential cha-
racters of ZTrichloris.

9. Grmyorogon, Beauv. (Anthopogon, Nutt.), differs from all
the preceding one-flowered genera in the spikelets not closely
crowded, but more or less distant along the slender rhachis of the
spikes, although still sessile in two rows and unilateral ; the spikes
themselves are scattered or verticillate along the common pe-
duncle. There are four or five American species, northern or
southern, and one from Ceylon, G. rigidus, Thw., forming Nees’s
genus Dichetaria, but only differing from the American ones in
the spikes fewer in the panicle, and the spikelets rather larger with
longer awns. Doell’s G. foliosus and G. pullulans should be
restored to Chloris, with which they agree in every respect except
that the spikes are not quite so closely clustered at the end of
the peduncle.

10. Moyocuarz, Doell, a single Brazilian species of which I
have seen no specimen, is removed by Doell from Gymnopogon,
where Martius had placed it, as having no continuation of the
rhachilla beyond the flower. Nees, however, describes a bristle-
like continuation, but not bearing any empty glume or awn
as in Gymnopogon. The genus is as yet, therefore, in some
measure doubtful.

11. Scnrpornarpvs, Steud., is the. North-American Leptu-
rus paniculatus, Nutt., which, however, Steudel failed to recog-
nize. Nuttall indicated its affinity to Gymuopogon, and evidently
only placed it in Lepturus from not knowing the latter genus
except from the imperfect characters then published. Schedon-
nardus has now been figured in the Jast part of Hooker’s Icones ;
the description, accidentally omitted in printing, will appear in
the next part.

12. CraspeporHacuts, Benth., is a single species from east
tropical Africa, allied to Schedonnardus, but differing in the

104 MR. G. BENTHAM ON GRAMINEA.

flexuose rhachis of the spikes bordered by a narrow membrane,
in the flowering glume and palea vey small and thin, resembling
lodicules, and a few other minor points. It is being figured for
the forthcoming part of Hooker's Icones.

13. Bovrenova, Lag. (Hutriana, Tein., Actinochloa, Willd.),
comprises about twenty-five American species, northern or
southern, but chiefly western. As in the four preceding genera,
the spikes are distant along the main peduncle, and often nume-
rous, very rarely reduced to one or two; but they are usually
short, with the spikelets densely crowded in two rows on one side
of the rhachis, and the raachilla always continued beyond the
single hermaphrodite flower, bearing one to three empty glumes
or awns, or sometimes a male flower. The flowering and upper
empty glumes usually end in three or five lobes, points, or awns;
but they are often exceedingly variable in this respect even in
the same specimen, and it becomes difficult to make much use
of them in the arrangement of the species. The following four
sections, raised by some to the rank of genera, are founded chiefly
on inflorescence :—(1) Chondrosia or Chondrosium, Desv. Spikes
usually few, often rather long, with numerous spikelets (more
than twelve) neatly pectinate, and the terminal empty glume
usually three-awned ; the species rather numerous, especially in
Mexico, where they run much one into another. (2) Atheropo-
gon, Muehl., including Heterostega or Heterosteca, Desv. Spikes
often numerous, but usually very short with few (rarely above
twelve) spikelets, crowded but scarcely pectinate, or almost re-
duced to clusters, the terminal empty glume varying from three-
awned to entire, or reduced to a single bristle. The species best
known, B. racemosa, Lag. (Atheropogon apludoides, Muehl.,
Dinebra curtipendula, DC.), was associated by DeCandolle and
Beauvois with the Dinebra arabica of Jacquin, which, however,
differs essentially in its several-flowcred spikelets. (8) Tria-
thera, Desy. Spikes still further reduced than in Atheropogon,
consisting usually of two to four spikelets so narrow and so close
together as to appear like a single one, and perhaps sometimes
really only a single one, the upper empty glume reduced to three
awns, as in several species of the preceding section. There appear
to be two species, B. aristidoides (Dinebra aristidoides, H. B. K.,
forming the genus Aristidiwm, Eudl.), with two to four spikelets
to each spike, and B. tréathera, to which I should, with Munro,
refer both Triathera, Desv., and Triana, H.B,K. The spike-

MR. G. BENTHAM ON GRAMINES. 105

lets vary in the same panicle, one, two, or three to the spike,
and are themselves polymorphous. Where there are three, I
have found the lowest empty glume of the lowest spikelet very
narrow and awnlike, and very probably that which Kunth has
described and figured as an awn at the base of the glume; in
the uppermost of the three spikelets the lowest empty glume is
similar to the second, the intermediate spikelet being sometimes
like the upper, sometimes like the lower one. (4) Polyodon,
H. B. K. (Zriplathera, Endl.). Spikes few, short, and crowded at
the.end of the peduncle with few spikelets, the flowering glume
three-awned, the two or three upper empty glumes each with
three or five awns, having together the appearance of a single
cluster of many awns. We have two species, B. disticha (Poly-
podon distichum, H.B.K.), including apparently <Atheropogon
afinis, Fourn. (not Eutriana affinis, Hook. f.), with the spikelets
including the awns under half an inch long, and B. multiseta (Eu-
triana multiseta, Nees), with the spikelets and their awns above
an inch and a half long, giving the plant the aspect of Boissiera.
Corethrum, Vahl, is probably the same plant. He received his
specimen in a collection of Syrian plants sent him by Thouir,
into which it had probably got misplaced by some carelessness in
sorting. No special locality is ascribed to it nor any indication
of the collector; and no plant answering to Vahl’s elaborate and
probably accurate description is known from Syria or auy part of
the Levant.

14. Mrzanocencaris, Nees, comprises three species from
East India or tropical Africa, closely resembling each other, and
at first sight having the aspect almost of yopogon but the
characters are very nearly those of Bouteloua (Atheropogon);
the genus is readily distinguished from both by the linear plumose
empty glumes.

The preceding genera have all only a single flower in the
spikelet, the second, when present, being male only ; in the follow-
ing ten genera there are at least two, and often several more
fertile flowers.

15. Trirogon, Roth (Plagiolytrum, Nees), contains about eight
East-Indian and tropical-African species, with the single elon-
gated terminal spike of Enteropogon, but with several-flowered
spikelets, and the flowering glumes more or less three-awned as
in Trichloris, Triraphis, etc., the lateral awns sometimes reduced
almost to teeth.

106 MR. G. BENTHAM ON GRAMINEA.

16. Leprpoprronia, A. Rich., is a single Abyssinian species
which I have not seen, and of which the specimen described is
said to have been imperfect. ‘From the figure and description it
would appear to be allied to the Zripogon abyssinicus, Nees, but
with broad, very villous flowering glumes, and the single awn not
quite terminal. :

17. Tetrapogon, Desf., has four Abyssinian, North-African,
or West-Asiatic species, including the above-mentioned Chloris
villosa, Pers., and C. macrantha of Jaubert and Spach. They have
one, two, or rarely three terminal erect spikes, resembling Elio-
nurus in the long silky hairs which cover them, but with the cha-
racters of Chlorides, differing from Chloris itself in their several-
flowered spikelets.

18. AsrreBia, I. Muell., comprises two or three Australian
species formerly referred to Danthonia, from which the habit and
untwisted awn separate them. In the ‘Flora Australiensis’ I
placed them near Pappophoree on account of their many-nerved
glumes; but the inflorescence places them in Chloridez, where
they come in many respects near to Letrapogon.

19. Waneunnemuta, Moench (Cynosurus Lima, Linun.), from
Spain and North Africa; 20. Crenopsis, DeNotar. (Festuca
pectinella, Delile), from North Africa; and 21. Terracuye,
Nees, from South Africa, are all single species hitherto referred
to Festucere ; but tliey have all the one-sided spikes with the
spikelets sessile in two rows of Chlorides, to which, following out
Munro’s memoranda, I have transferred them. The spikes are
solitary, erect, and often slightly faleate in Wangenheimia and
Ctenopsis, several scattered along the common peduncle in
Tetrachne.

22. Divepra, Jacq., remains limited to the original African
and East-Indian D. arabica figured by Jacquin. DeCandolle, as
above mentioned, joined it with the section Atheropogon of
Bouteloua, of which it has the habit; but there are always two
fertile flowers to the spikelet. Kunth referred it to Leptochloa,
from which it is further removed by its short dense awned spikes ;
and from both it is separated by the lower empty glumes as long
as, or longer than, the rest of the spikelet,

23, Exzvsine, Gertn., taken in the sense given to it by Per-
soon, isa natural genus of about seven species, from the tropical
and subtropical regions of the Old World, two of them common
weeds also in America, The spikes are usually several, digitate

MR. G. BENTHAM ON GRAMINER. 107

at the end of the peduncle, and in some species with the addition
of others scattered or verticillate lower down. The flat spikelets
have been sometimes mistaken for those of Eragrostis ; but their
arrangement in two rows is always that of Chloridee. The genus
is often restricted to Gertner’s EZ. coracana and E, indica, in
which the spikes are digitated and rather long, and the membra-
nous pericarp loose on the ripe seed. This character is particu-
larly marked in the Z. coracana; but that is probably a plant
somewhat modified by long cultivation. Inthe common Z. indica
the pericarp is often as loose, but sometimes remains very thin
and not so easily detached. In the still more common E. egyp-
tiaca, Pers., forming the genus Dactyloctenium, Willd., the digi-
tate spikes are very short and dense, and the very thin pericarp
appears to wither away or to dry up in ripening, leaving the seed
apparently exposed and rugose, similar to that of E. indica. In
E. brevifolia, Wall. Cat., and E. glaucophylla, Munro (Dactylo-
ctenium glaucophyllum, Courb.), the spikes are short as in F.
egyptiaca, and more or less of the remains of the membranous
pericarp may be often seen persistent about the seed. Acrachue,
Wight and Arn., is the #. verticillata, Roxb., in which the spikes
are rather long as in Z. indica ; but besides the terminal digitate
ones, there are others scattered or verticillate along the peduncle.
Arthrochlena, Boiv. in herb. J. Gay, is a remarkable Madagascar
species which may be thus defined :— 2. macrostachya, Benth.,
elata, foliis angustissimis rigidis crassiusculis, spicis 2-3nis ter-
minalibus, spiculis confertis 18-20-floris, glumis acute carinatis
paleaceis bifariam imbricatis. A plant of rushlike habit about
2 feet high, the spikes about 4 inches long, with very numerous
spikelets varying from 4 to 6 lines long, resembling those of Hra-
grostis, but much more rigid.

24, Luprocutoa, Beauv. (Oxydenia, Nutt.), about twelve spe-
cies, tropical or subtropical, in the New as well as the Old World,
and extending on the one hand into North America, and on the
other into extratropical Australia, is one of those genera which
interfere provokingly with our classifications. Nearest allied to
Eleusine, it has also considerable affinity with Cynodon, Diplachne,
and Poa, to which some of the species have been occasionally
referred; and one bas been figured as a Cynosurus. The chief
character consists in the slender spikes scattered along the
common peduncle, with numerous small flat spikelets, giving the

108 MR. G. BENTHAM ON GRAMINER.

plants a different habit from that of the several preceding genera.
The species may be distributed into two rather distinct sections.
(1) Pseudocynodon, with only one or two flowers in the spikelet, dif-
fering but little from Cynodon except in inflorescence. To this sec-
tion belong L.wniflora, Hochst.( Cynodon gracilis, Nees), from Abys-
sinia, DL. Neesti (Cynodon Neesii, Thw.), from Ceylon, and L. poly-
stachya, Benth. (Cynodon polystachyus, Br.), from Australia. (2)
Euleptochloa, with two cr more flowers to the spikelet, comprises
the remainder of the species. Those which have a point or short
awn to the flowering glume were formerly generically separated
by Beauvois under the name of Rabdochloa. Amongst the pub-
lished species, Z. arabica, Kunth, is the genus Dinebra; L. Lind-
leyana and ZL. mollis, Kunth, are referable to Triodia. L. dubia,
Nees (Chloris dubia, H. B. K.), and the North-American LZ. (Di-
plachne) fascicularis, A. Gray, appear both to be true species of
Diplachne. Lorentz’s South-American plant distributed by Gri-
sebach as L. fascicularis appears to be amere variety of Panicum
sanguinale. LL. Wightiana, Nees, is an Eragrostis. L. plu-
mosa, Anders., is a Triodia.

There remain three anomalous monotypic genera from the
Mexican-Texan region—25. Bucu1ok (Sesleria, afterwards Oalan-
thera, Nutt.), 26. Jouvea, Fourn., and 27. Optzra, Pres!, which
connect in some measure Chloridez with the subtribe Sesleriea,
and are all dicecious, and very remarkable for the great dissimi-
larity in the spikes and spikelets of the two sexes. Of the first,
numerous specimens fully confirm Engelmann’s excellent figures
and description; the other two are unknown to me, and remain
somewhat doubtful. Of Jowvea, Fournier only knew the female,
which he says is allied to Buchioe, with a very different habit.
Opizia is said by him to have the male plant exactly like Buchloe ;
the female figured by Presl must be very different, though his
and Fournier’s descriptions do not agree in all points. ‘

Tribe XII. Festtcem,

The large tribe Mestucee presents considerable difficulties to
the systematist. Of the seventy genera on our list (about a
hundred and ten of some botanists), the greater number are perhaps
better defined than those of Agrostez for instance, and afford a
much greater variety of characters; but none of the various
arrangements’ proposed for distributing them into groups or sub-
tribes have proved satisfactory, and the two largest genera Poa

MR. G, BENTHAM ON GRAMINER. 109

and Festuca are connected by a number of smaller ones, which
are more variously associated together or separated by European
botanists than almost any others of the Order. As a whole,
Festucez should include all the Poacee with two or more perfect
flowers to the spikelet, which have neither the peculiar inflores-
cence of Chlorides or of Hordeez, nor the dorsal or twisted awn of
Avene, nor the peculiar habit of Bambusex. But we have seen
that there are a few species where the awn is wanting, but which
must yet be left in Avenee; we shall find that in Diplachne, Oreo-
chloa, and even in Festuca itself, there is occasionally an inflores-
cence very nearly that of Chlorides ; and with regard to Bambusee,
distinct as is the habit and foliage of the great mass of genera, yet
it is exceptional in Planotia, and in the subtribe Centotheceex of
Festucee there is some approach to that of the Bambusew. The
subdivisions proposed of the Order into subordinate groups
have been so various, and often on such plausible (though some-
times contradictory) grounds, that it is not without hesitation
that I have selected for adoption the following eight subtribes :—

Subtribe 1. Pappophoree, has often been raised to the rank of
a substantive tribe, but with various limits; and it really is only
distinguished from Triodiew by the more numerous teeth, lobes,
or awns of the flowering glumes. There are-five well-established
genera, requiring little comment.

1. Pommerrvtia, Linn. f., is a single East-Indian annual,
with short spikes almost enclosed in the upper leaf-sheaths, and
remarkable for the presence of two empty glumes between the
ordinary lower pair and the flowering ones, as in Ctenium, Ere-
mochloa, Brylkinia, and Uniola.

2. PappopHorum, Schreb., has nearly twenty species from the
warmer regions of both the New and the Old World, distributed
in two sections, regarded by some as distinct genera :—Poly-
rhaphis, Trin., a few American species in which the flowering
glume has thirteen to twenty-three very unequal awns; and
Enneapogon, Desv., chiefly, but not exclusively, from the Old
World, in which the flowering glume has nine awns, all nearly
equal, or five of them rather external and slightly different from
the four inner ones.

8. Correa, Kunth, is a single tropical American species,
differing from Pappophorum in the looser panicle, and in the
flowers usually more than two, instead of only one or two, in
each spikelet.

LINN. JOURN.— BOTANY, VOL, XIX. K

110 MR. @. BENTHAM ON GRAMINER.

4. Borsstrra, Hochst., is a single Oriental species, which,
besides the characters derived from the glumes, has a very short
membranous dilated two-lobed style, different from that of every
other known grass.

5. Scumiptra, Steud., contains two species closely resembling
each other, one from the Cape-Verd Islands, the other from
tropical and South Africa, with a narrow but loose panicle, and
the flowering glumes ending in four inner lanceolate lobes, and
tive outer subulate lobes or awns.

Our second subtribe, Triodiee, is not so definite as could be
wished. There are usually more than two flowers to the spike-
let; and the flowering glumes have rarely more than three nerves,
and end in three teeth, lobes, or awns. These characters are gene-
rally very prominent; but in a few species of Zriodia and Dipl-
achne the teeth are scarcely more than what occur occasionally
in some other subtribes, and in one or two species of Triraphis
the nerves of the glumes are more numerous, bringing them
technically near to Pappophorew. The panicle in all the genera
is usually narrow, dense or loose, but very rarely spike-like, and
in a few species loose and spreading. We include six genera.

6. Trropra, Br. (Uralepis, Nutt.), as at present limited, com-
prises about twenty: extratropical species, northern or southern,
with a very few extending into the tropics in America or Africa.
It has the typical characters of the tribe without the peculiarities
of the other genera, the lobes of the flowering glumes reduced to
short teeth or points, or the central one rarely lengthened into a
short awn. It must be admitted, however, that it is still both a
vague and a polymorphous genus, comprising three rather dis-
tinct sections and a few anomalous isolated species :—1. Isotria,
three Australian species (I. Mitchelli, Benth., I. pungens, Br.,
and J. Cunninghamii, Benth.) with the three lobes of the flower-
ing glumes narrow lanceolate and equal; 2. Uralepis (Sieglingia,
Bernh., Merisachne, Trin.), about sixteen American or Australian
species with one European one, in which the lateral teeth of the
glumes are broad and not pointed and sometimes very minute,
the middle one a point or very short awn; 3. Zricuspis, Beauv.
(Windsoria, Nutt.), three North-American species differing from
Uralepis in the nerves of the lateral teeth produced into short
points. Besides these, Leptocarydion, Hochst., is an Abyssinian
species, 7. plumosa (Leptochloa plumosa, Anders., Diplachne alo-
pecuroides, Hochst.) with the dense soft panicle almost of Zri-

MR. G. BENTHAM ON GRAMINER, 111

chloris, but with the spikelets of Zriodia. Trichoneura, Anders.,
is Leptochloa Lindleyana, Kunth, from the Galapagos Islands,
with the habit nearly of Diplachne fascicularis, but with the
characters of Triodia. Leptochloa mollis, Kunth, from Senegal,
which I have not seen, would appear from his figure to be a
Triodia, near to the ZL. plumosa, but with the loose panicle of
T. Lindleyana. Rhombolytrum, Link, is T. filiformis, Nees, a
Chilian species, very near to the North-American T. albescens,
and . trinerviglumis, Munro, and also to 2. Kerguelensis,
Hook. f., and 7. antarctica (Catabrosa antarctica, Hook. f.), from
extreme southern America.

7. Dretacuye, Beauv., now comprises about fourteen species,
dispersed over the tropical and temperate regions both of the
New and the Old World, and variously referred to Triodia,
Leptochloa, or Molinia by different agrostologists ; and the genus
is really closely connected with the two first-named, but more
especially with the Zriodie of the typical section Uralepis.
From these it chiefly differs in inflorescence: the branches of
the panicle are long and slender; the spikelets, almost linear,
scattered along the rhachis and sometimes sessile or nearly so
in two rows, but not regular and unilateral enough to place the
genus in Chloride, to which it is sometimes referred. The cha-
racteristic teeth of the flowering glumes are also sometimes very
minute. Whatever position, therefore, we give to the genus, it
must be more or less an arbitrary one; but that next to Triodia
seems the least objectionable.

8. Tripnasts, Beauv. (Diplocea, Rafin.), has two North-Ameri-
can species (Uralepis cornuta, Hll.,and U. purpurea, Nutt.), with
a narrow, slender, slightly-branched panicle, and the flowering
glumes deeply divided into three narrow lobes, the central one a
slender awn. The South-American Triplasis setacea, Griseb., is
a Dipiachne (D. spicata, Doell).

9. Scteroroaon, Philippi (Lesourdia, Fourn.), has four species,
one from Chile, the others from the Mexican-Texan region, all
remarkable for the unisexual spikelets, those of the two sexes so
different in aspect that without positive evidence it would have
been difficult to suppose them to belong to the same plant. The
Mexican ones have been very well described and one of them
figured by Fournier, who, from his specimens, supposed them to
be strictly dicecious ; but we have specimens with the two in-
florescences upon different branches of the same individual.

K2

112 MR. G. BENTHAM ON GRAMINES.

Fournier also was not aware of the Chilian species previously
published by Philippi, whose generic name necessarily takes the
precedence over Fournier’s.

10. Eremocutos, 8. Wats., contains two New-Mexican or
Californian species, well described and illustrated by S. Watson.
The genus is exceptional in the subtribe in its one-flowered
spikelets, and in the two empty three-awned glumes between
the lower pair and the flowering one.

11. Tarrapuis, Br., comprises five or six species, one as yet
unpublished from South Africa, the others from Australia. In
all, the three lobes of the flowering glumes are awned, whilst in
one species, Z. mollis, Br., otherwise inseparable from the genus,
there are additional small membranous lobes between the awns,
and at least five nerves to the glume, showing a close connection
with the Pappophoree. Z. microdon, Benth., from Australia, is
a very anomalous plant with unawned glumes, which should be
removed from the genus; but itis difficult to say where it should be
placed 1. capensis, Nees, is a Danthonia (D. radicans, Steud.).

The third subtribe, Arundinee, though often regarded as a
substantive tribe, has no definite character beyond the tall habit
with a rich panicle, as designated by the common name Reed,
and the long hairs surrounding the flowering glumes, either
arising from the rhachiila or from the glumes themselves. This
character, however, is no more exclusive here than in Saccharee ;
for there are other genera, such as Culamagrostis, Graphephorum,
&c., which have the hairs nearly as long, but which on other
accounts cannot be included in Arundinee. The genera certainly
belonging to the subtribe are four:—12. Gynerium, Humb.
and Bonpl., three tropical or subtropical American species with
strictly dicecious spikelets. 18. AmprnopEsmos, Beauyv., two
Mediterranean species with rigid five-nerved flowering glumes.
14. Arnunpvo, Linn. (Donax, Beauv., Scolochloa, Mert. and Koch,
Amphidonax, Nees), three or four species from the warmer
regions of the New and the Old World, very abundant all round
the Mediterranean, with three-nerved flowering glumes ; and 15.
Punraemires, Trin. (Arundo, Beauv., Czernya, Presl, Trichodon,
Roth), two species, one of them almost cosmopolitan, only differ-
ing from Arwndo in the lowest flower of the spikelet being male
only. These last three genera are frequently, and perhaps with
reason, regarded as sections only of Arundo. Two Mexican
monotypic genera proposed by Fournier, 16. Gournta, and 17.

MR. G. BENTHAM ON GRAMINES. 113

CaLamMocHioa, are unknown to me. They are placed by the
author in Arundinee ; but both would appear, from the short
character given, to have a rather different inflorescence.

We have formed our fourth subtribe Sesleriee of ten genera,
all except the monotypic Elytrophorus extratropical, and con-
nected together for the most part by inflorescence as well as by
structural characters. These are, however, not without excep-
tions. The spikelets are usually collected together in little heads
or close clusters, which are themselves closely clustered ina dense
globular or spike-like panicle; and at the base of the head or
panicle are usually a few barren spikelets or empty glumes, or,
sometimes single glumes, which I have elsewhere, though as I
now believe erroneously, designated as bracts subtending the
branches of the inflorescence. In two genera the whole inflo-
rescence is reduced to two or three spikelets sessile in a cluster
of floral leaves. In structure the flowering glumes are variable
in their nerves; but the styles, in almost all the genera, are long,
with barbellate or very shortly plumose stigmatic branches,
forming an exception to the tribe, which has induced several
botanists to refer Sesleria itself to the Panices, from which it
differs so widely in other characters. We include the following
genera in the subtribe :—

18. MowantHocutoz, Engelm., is a single Texan species, ano-
malous in habit and character, well described and figured by
Engelmann. It has been compared to Buchioe on account of its
unisexual spikelets and creeping habit; but the two sexes in
Monanthochloe are very similar to each other, and there is no
indication in the inflorescence of any affinity with Chlorides.

19. Munroa, Torr., has now three or four species, two or three
from extratropical South America having been recently added to
the original Mexican-Texan one, figured in the last part of Hooker’s
Icones. The genus is a perfectly isolated one, showing only
some slight affinity with Monanthochloe, especially in the very
few spikelets sessile within a cluster of floral leaves; they are not,
however, unisexual as in that genus.

20. Ecuinaria, Desf. (Panicastrella, Moench), a single well-
known Mediterranean species ; 21. AMmocHLOA, Boiss. (Cephalo-
chloa, Coss. & Dur.), two Oriental or North-African species ;
and 22. Urocuiana, Nees, a single South-African species figured
in the last part of Hooker’s Icones, require no further comment
on the present occasion.

114 MR. G. BENTHAM ON GRAMINEA.

23. Sesterta, Scop., about eight European or West-Asiatic
species, is an old-established distinct genus which has been but
little interfered with, except that S. tenella, Host, has been pro-
posed as a genus by Link under the name of Psilathera, but
upon characters which do not appear to be of more than specific
value.

24, Etyrropuorus, Beauv. (Echinolysium, Trin.), is a single
exceptionally tropical species, widely spread, but limited to the
Old World. The little heads of minute spikelets forming a spike
longer than usual in the subtribe, and often interrupted, and the
wings on one or both the keels of the palea readily distinguish
the genus.

25, Finerruutuia, Nees, a single species figured in the last
part of Hooker’s Icones, is exceptional in the whole primary
division Poacew in having the very short pedicels articulate
below the spikelet as in Panicacez, whilst the male flower or
empty glume is above the fertile flower asin Poacew. Its geo-
graphical range is also peculiar: rather common in South Africa,
it bas been recently gathered by Dr. Aitchison in Afghanistan,
without its ever having been observed in any intermediate station.

All the preceding genera have the long styles of the subtribe ;
but there remain two with the styles scarcely longer than in the
other Festucee, whilst the barren spikelets at the base of the
dense inflorescence or of its branches are very conspicuous, and
show a close affinity with the Sesleriew. These are 26. La-
MaRCKIA, Moench (Chrysurus, Pers., Pterium, Desv., Tinea,
Garzia), a single Mediterranean species, and 27. Cynosurvs,
Linn., three or four species with a much wider geographical

range over the temperate regions of the Old World, and one of | '

them at least now naturalized in several other countries within or
without the tropics. Both genera are remarkable for the lower
barren spikelets of the clusters or spikes elegantly pinnate with
numerous bifarious empty glumes. The two were united by the
older botanists, and have been again brought together by some
modern ones under the Linnean name Cynosurus ; but they appear
to be sufficiently different in habit and character to be maintained
as separate genera ; and Cynosurus itself has two very distinct sec-
tions, raised by some to the rank of genera :—Cynosurus proper
for the C. eristatus, Linn., and its annual Algerine variety 0.
polybracteatus, Poir., altered to C. multibracteatus, Roem. & Schult.
(C. erista-galli, Munby ), in which the spike-like unilateral panicle

MR. G. BENTHAM ON GRAMINER. 115

much resembles the spikes of Chloridee, and the glumes, though
very pointed, are unawned; and Fulona, Adans., altered by Du-
mortier to Phalona, for the C. echinatus, Linn., and C. elegans,
Desf., with the panicle or head more like that of a Dactylis, but
with awned glumes.

The characters of our fifth subtribe, Hragrostee, like those of
Eufestuces, are chiefly negative. The two together comprise all
the Festucew which have not the peculiarities of either of the
other six subtribes, and differ from each other in the Eragrostee
having three prominent nerves to the flowering glumes, the Eufes-
tuces five or more nerves, sometimes rather obscure. Trifling as
this character may be, it is a fairly constant one, the exceptional
species being exceedingly rare ; and I have found no other one
so useful in distributing these numerous genera into groups.
We have included in Eragrostez twelve of them, though the last
of them (£etrosia) might be equally well placed under the fol-
lowing subtribe Melicez.

28. Karta, Pers., has about twelve species, of which ten
are European, temperate Asiatic, or North African, one of them
extending over extratropical America north and south and South
Africa; the two remaining species are endemic, one in South
Africa, the other in the Sandwich Islands. The genus has been
generally admitted with little variation ; but it is difficult to assign
to it any positive character. The panicle is usually dense and
narrow, often spike-like; andthe glumes are more scarious, espe-
cially on the margins, and have fainter nerves than in the others
of the subtribe. It has been divided into two sections, maintained
by some as genera:—1l. Airochloa, Link, to which Reichenbach
restricts the name of Aeleria, with the glumes obtuse or acute
but without distinct points; and 2. Lophochloa, Reichenb., to
which Link restricts the name of Keleria (digialitis, Trin.,
altered by Schultes to yzalina), in which the flowering glumes
have a distinct point or short awn at or just below the tip. This
section includes, besides the species enumerated by Cosson and
Durieu in their ‘ Flore d’ Algérie,’ K. Gerardi, Munro, from South
Africa, and K. glomerata, Kunth (K. vestita, Nees), from the
Sandwich Islands. The last species differs slightly from the
genus in the long loose panicle, which, however, is more dense in
our specimens than it is figured by Kunth. C. Koch’s genus
Withelmsia, from the Caucasus, is, according to Grisebach, only
a depauperated specimen of K. phleoides, Pers.

116 MR. G. BENTHAM ON GRAMINED.

29. AVELLINTIA, Parlat., isa single West-Mediterranean annual
with the habit of Schismus, and placed by Saviin Bromus, by
Gussone in Avena, and by DeCandolle in Keleria. It is well
marked amongst Eragrostee by the outer glumes, of which the
lowest is almost reduced to a bristle, and the second broad, mem-
branous, and the largest of the spikelet; the flowering glumes
awned.

30. Earonta, Rafin. (Reboulea, Kunth, Colobanthus, Trin.),
two or three closely allied North-American species, with the
second empty glume the largest of the spikelet, as in Avellinia;
but the habit is very different and the glumes all unawned.

81. Dissanrnerium, Trin. (Phalaridium, Nees, Stenochloa,
Nutt.), comprises two, or perhaps three, species from the Andes
of South America and the coasts of Mexico and California,
figured and described in the last part of Hooker’s Icones. They
have most of the characters of Schismus, but, besides the widely
distant geographical stations, they differ in the nerves of the
flowering glumes always three, not five.

82. Moxrn1a, Meench (Enodium, Gaudin), a single well-known
European and temperate Asiatic species, and 33. SpHEnopvs,
Trin., a very pretty little Mediterranean annual, require no fur-
ther comment on the present occasion.

34. CaraBrosa, Beauy., can only be distinctly characterized if
reduced to the single C. aquatica, Beauv., placed by some authors
in Aira, by others in Glyceria. In it the three nerves of the
flowering glume characteristic of Eragrostee are very promi-
nent. The two or three Oriental species added to it by Trinius
belong to the genus Comodium. C.antarctica, Hook. f., is a Triodia.
C. glaucescens, Phippi, and C. magellanica, Hook. f., are true
Glyceria.

35. Eragrostis, Beauv., an almost cosmopolitan genus of
above eighty species (multiplied by Steudel and others to about
two hundred and fifty), is a very natural one so far as the great
majority of species are concerned, and distinctly limited if we
include the three-nerved glumes amongst the essential characters.
Yet in other respects there are exceptional species which have
been variously referred, even by modern botanists, to Poa, Fes-
tuca, Briza, Dactylis, Eleusine, or Leptochloa, which they in some
measure approach respectively ; and some have been proposed as
substantive genera; but it has appeared to me that the genus
may be best defined if retained entire, dividing it into the six

MR. G, BENTHAM ON GRAMINER. 117

following fairly distinct sections :—1. Cataclastos, Doell, includes
E. ciliaris, Link, EZ. peruviana, Trin., and a few other tropical
species, which, on account of their short spikelets with few
flowers and fragile rhachilla, have been restored to Poa by Four-
nier and some others; but the shape and nervation of the glumes
are quite those of Hragrostis, and the inflorescence, though pecu-
liar, is not more that of Poa than of Hragrostis. Macroblepharus
of Philippi, judging from his description, does not seem to differ
from the true ZH. ciliaris. 2. Plagiostachya comprises some
African and East-Indian species, which, with the flat several-
flowered spikelets and continuous rhachilla of Eragrostis proper,
have an inflorescence approaching that of Chlorides. The species
are rather dissimilar in habit. #. bifaria, Steud., has the long
simple terminal spike nearly of Tripogon, with obtuse glumes.
E. Schimperi (Harpachne Schimpert, Hochst.) has a shorter simple
spike and acuminate flowering glumes. J. brevifolia and £.
Oclachyrum, Benth., the latter forming the genus Calachryum,
Nees, and figured in the last part of Hooker’s Icones, have
nearly the habit of some species of Eleusine (Dactyloctenium) or
of Hluropus. E. congesta, Oliv., and EH. cynosuroides, Roem. et
Schultes, have very numerous short sessile spikelets crowded or
clustered along the long terminal peduncle. 38. Myriostachya is
an East-Indian species, H. Wightiana (Leptochloa Wightiana,
Nees), allied to £. cynosuroides, but with a more complicated in-
florescence. 4. Pteroéssa, Doell, or Eragrostis proper, is charac-
terized by the usually many-flowered spikelets with the rhachilla
continuous or rarely articulate when old, the flowering glumes
usually deciduous, leaving the palea persistent on the minute
floral axis with its back to the rhachilla. The species are nume-
rous, and may be distributed into three rather distinct series :—
Cylindrostachy@, three or four Australasian species, with narrow-
linear almost terete spikelets ; Leptostachye, including the cos-
mopolitan Z. pilosa, Beauv., and its allies, with narrow-linear flat
spikelets;. and Megastachye, including the widely-spread E. mega-
stachya, Link, and many other, chiefly American, species, with
broader linear or oblong flat spikelets. The generic name Mega-
stachya has undergone many vicissitudes. It was first founded by
Beauvois on the Poa mucronata of his Flora of Oware and
Benin; but in drawing up the generic character for his ‘ Agro-
stographie’ he had in view chiefly the common FE. megastachya.
Fournier, more recently, founded a genus on those American

118 MR. G. BENTHAM ON GRAMINER.

species (Z. reptans &c.) which have the spikelets more or less
unisexual, the males usually flatter, longer, and with more flowers
than the females; but this separation of the sexes is very
variable, and not always accompanied by any difference in habit.
Some species may be occasionally quite dicecious; in others the
males and females are in different panicles on the same plant ;
others, again, are variously polygamous; and here, as in the
Chilian Pow, the character is too inconstant to justify generic
or even sectional separation. Jlegastachya, Fourn., does not in-
clude #. megastachya, Beauv. 5, Platystachya, includes a few
African or Arabian species with broad, flat, many-flowered spikelets
with rather paleaceous glumes, and the rhachilla articulate as in
Cataclastos. Amongst these Munro would include as 2. genicu-
lata the Briza geniculata, Thunb., which in its thickened spikelets
appears intermediate between Briza and Eragrostis, but has the
prominently three-nerved glumes of the latter. 6. Sclerostachya,
has three African or Asiatic species. The paleaceous glumes and
articulate rhachilla are those of Platystachya; but the spikelets
are not so broad, and the rigid habit with long and rush-like or
short and pungent leaves are those of 4luropus.

36. Ipyum, Philippi, is a single Chilian species unknown to
me. From the author’s description, it would appear to differ
from the section Pteroéssa (Cylindrostachye) of Eragrostis in the
articulate rhachilla and in inflorescence.

387. Curanpa, Willk., is a genus proposed for the European or
North-African Festuca maritima, DC., F. philistea, Boiss., F.
memphitica, Boiss., F. divaricata, Desf., F. incrassata, Salzm., and
F. lanceolata, Forsk., which, notwithstanding a general resem-
blance to Festuca in the shape of the spikelets, could not remain
in that genus without an essential alteration in its generic cha-
racter, the glumes being strongly keeled and three-nerved from
the base. The inflorescence is also peculiar. ,

38. Orzocuxoa, Link, is a single European mountain species,
formerly included in Sesleria, which it approaches in its short
compact inflorescence and the slightly elongated stigmas; but
there are no barren spikelets, the spike is simple, with almost
sessile bifarious spikelets like those of some species of Festuca
(Scleropoa), and the glumes are those of Eragrostes.

39. Ecrrosta, Br., comprises three or four Australian species
allied both to Eragrostez and to Melicexw, but technically rather
better placed in the former subtribe.

MR. G. BENTHAM ON GRAMINER. 119

As a sixth subtribe, Melicee, I have collected five genera
allied both to Eragrostex and Festucesx, but technically connected
with each other by their spikelets containing two or more empty
glumes above the flowering ones. This character is also observed
in Ectrosia and in Lophatherum, which I had formerly included
in the group; but on other accounts the former appears better
placed in Eragrostex, and the latter in Centothecee.

40. Cryprocutoris, Benth., a single species probably from
Patagonia, 41. HerEracune, Benth., two Australian species, and
42, AnrHocHLos, Nees, one or two species from the Andes of
South America, are very distinct dwarf grasses, described and
figured in Hooker’s Icones.

43. Meuica, Linn., contains above thirty species dispersed over
the temperate regions of the northern hemisphere, and extending
down the Andes into extratropical South America, represented
also in South Africa, This genus, the typical representative of
the subtribe, has, been universally recognized since the days of
Linneus, and less tampered with than any other genus of equal
extent, although it may be in some degree polymorphous in habit
as well as in character. In the typical delicas, however varied
the panicle, long and narrow, or very loose and spreading, the
spikelets are generally nodding or pendulous, with rarely more
than two flowers ; the flowering glumes more or less scarious on
the margins and never awned, and the terminal empty glumes
one within the other, form an obovoid obtuse mass. Jna section
proposed by Thurber for four North-west American species under
the name of Bromelica, the spikelets are erect, with more rigid
glumes occasionally awned and three to eight flowers, the upper
empty glumes narrower and not so closely packed, giving the
plants altogether so different an aspect, that I have much hesi-
tated whether I should not, as suggested by Thurber, raise the
section to the rank of a genus. In both sections the flowering
glumes have five or more nerves. J. stricta, Boland., is in some
measure intermediate between the two. Chondrachyrum, Nees, a
Brazilian species which I have not seen, would seem from the de-
scription given to be a true Melica.

44. Diarruena, Rafin. (Korycarpus, Corycarpus or Remeria,
Zea, Onoea, Franch. and Sabat.), two species, one from North
America, the other from Japan, is very near Melica; but the
flowering glumes have only three nerves and are hardened round
the grain, which usually exceeds them, and the stamens are reduced

120 MR. G. BENTHAM ON GRAMINEE.

to two or one. The habit is nearly that of the section Bromelica
of Melica.

Our seventh subtribe Centothecee is formed of a small number
of tropical grasses, several of which have been occasionally re-
ferred to Bambusez, but expelled from that tribe by all who have
specially worked at it. The structure of the spikelet is that of
some Eufestucee or Melicee; but the foliage is unusual, the
lamina of the leaf is broad and flat, and between the numerous
longitudinal veins are small transverse veinlets not observed in
any others of the Order except ina few Bambusex. There is,
however, none of that articulation of the lamina on the leaf-sheath
which is almost universal in the latter tribe. The Centothecem
comprise five genera.

45. Ceytornzca, Desv., has two or three species from the tro-
pical regions of the Old World. They are tall grasses with a
loose panicle, the spikelets awnless with usually more than two
flowers, without any, or with only one empty glume above them.
In the common C. lappacea, Desv., the flowering glumes have on
their back a few reflexed rigid hairs or bristles; and that has
been generally relied upon as the essential character of the genus ;
but the bristles are sometimes reduced to one or two minute
tubercles, or even wanting, and in an African species (probably
the Poa mucronata, Beauv.) there is no trace of them, and yet
the plant is in all other respects an undoubted Centotheca.

46. OrrmociaDa, Beauv., is asingle tropical-American species,
with the habit, foliage, and inflorescence of Centotheca ; but the
spikelets contain only one, or rarely two, fertile flowers, the second
flower being usually male only ; the glumes have never the re-
versed bristles of Centotheca; and the spikelets appear to be fre-
quently unisexual.

47. Lopuarurrum, Brongn. (Acroelytrum and Allelotheca,
Steud.), has one, or perhaps two, species from tropical and
Eastern Asia. The habit, foliage, and inflorescence are those of
Centotheca and Orthoclada; but above the single fertile flower
are several small empty, very shortly-awned glumes, densely
crowded in a little unilateral tuft or crest, bringing the genus
into connection with Melicesx.

48. Srruptoerne, Beauv., is a single species sparingly scat-
tered over Hast India, tropical Africa and America, and the
southern states of North America, allied perhaps in some re-
spects to Lophatherum, but quite isolated in habit and a variety

MR. G. BENTHAM ON GRAMINEX. 121

of characters. The long narrow spikelets are few or numerous
in a long, rigid, terminal unilateral spike; and the exceedingly
long capillary styles become spirally twisted together far beyond
the glumes. There is a considerable variety in the number
of spikelets, in the number of flowers in each from one to
four; and even the stamens and styles are sometimes two, some-
times three ; but I have been unable to trace any connection of
these diversities either with each other or with geographical
stations so as to mark distinct species.

49. Zuverres, Schreb. (Senitis, Adans., Despretzia, Kunth,
Krombholizia, Fourn.), has five or six tropical-American species.
They have not all the tall habit of the preceding genera; for the
best-known West-Indian and Central-American species, origi-
nally described by Linneus under Apluda, is a much weaker
plant with smaller leaves; they are, however, broad and flat,
with the characteristic venation of the subtribe. The spikelets
in the genus generally have one fertile flower with two to five
male ones above it.

I have already adverted to our eighth and last subtribe Lufes-
tucee, or Festucee proper, as differing from Eragrostex in having
five or more nerves to the flowering glumes instead of three or
one only. Generally speaking, they have not the several-awned
glumes of Pappophoree and Triodiew, nor the barren spikelets nor
long styles of Sesleriew, nor the long hairs surrounding the
flowering glumes of Arundinex, nor the cluster of upper empty
glumes of Melicez, nor the peculiar foliage of Centothecex ; yet
there are here and there exceptional species showing an approach
to one or another of these characters, and interfering much with
any definite line of demarcation. We include in the subtribe
twenty genera, the last five of which are further characterized by
the adherence of the grain to the palea ; but, as already observed,
this character is not quite constant even in Festuca, and is occa-
sional in such genera as Poa and Briza, where the grain is
usually free. I have been unable to discover any other cha-
racter which would distribute the genera of the subtribe into
more satisfactory groups.

The first two genera have a simple racemose inflorescence.
50. Prevrorogoy, Br. (Lophochlana, Nees), has three species,
one arctic, the two others Californian, distinguished by the keels
of the palez bearing a linear tooth or flat crest; and 51. Brrt-
Kinta, F. Schmidt, a single Japanese species, with two empty

122 MR. G. BENTHAM ON GRAMINES.

glumes close under the flowering one besides the two lower per-
sistent ones.

52. Untona, Linn. (Zrisiola, Rafin., Chasmanthium, Link),
has four genuine North-American species, tall plants somewhat
variable in inflorescence, but all with flat broad spikelets in which
the three to six lower glumes are empty, but in size and shape
pass gradually into the fowering ones, which vary from three to
about twenty. If U. racemiflora, Trin. (U. virgata, Griseb.), from
the West Indies, be retained in the genus as having flat spikelets
with more than two empty glumes below the flowering one, it
must be considered as a very exceptional species with the inflo-
rescence nearly of Leptochloa among Chloridew. The small spike-
lets are closely sessile in two rows in unilateral spikes ; and these
spikes, shorter than in Leptochloa, are very numerous and crowded
along the long peduncle. It would be better perhaps to regard
the plant as a section of Leptochloa rather than as a distinct
genus. Fournier has added three Mexican species of Uniola
which are unknown to me; but, from his short characters, they
would scarcely seem to be true congeners. JU. prostrata, Trin.,
and its allies are now included in Distichlis.

53. Disricuis, Rafin., comprises four or five closely allied
species, or perhaps varieties ofa single one, extending from North
America down the Andes to extratropical South America, one of
them found also in Australia. They are generally, but not
always, maritime plants, with few spikelets nearly sessile in a
dense panicle, and generally if not always strictly diwcious, though
the two sexes differ but little in habit. The glumes are rather
rigid and paleaceous, which induced Link to join the only American
species with which he was acquainted, with the Mediterranean
Poa sicula, Jacg., as a genus Brizopyrwm, a name retained by
Pres] and by Fournier for the American species. The European
B. siculum, however, and some African congeners have the spike-
lets hermaphrodite, a more regular bifarious inflorescence, and
otherwise differ sufficiently from the American forms to be main-
tained as a separate genus which has a primary title to Link’s
name.

54. ALuropvs, Trin. (Oalotheca, Spreng., Chamedactylis, Nees),
has three species from the Mediterranean region, Central Asia,
and East India, formerly included in Dactylis, but differing in
their creeping or prostrate branching habit, short, rigid, ofien

MR. G. BENTHAM ON GRAMINER. 123

pungent leaves, more numerous flowers in the spikelets, and some
other minor points.

55. Daorytis, Linn., is now limited to two species :—the com-
mon and well-known D. glomerata, Linn., which from Europe and
temperate Asia has spread over many parts of the civilized world;
and D. cespitosa, Forst., the celebrated Tussock grass of the
Falkland Islands, which, though a much larger plant, appears to
be strictly a congener.

56. Lastocuioa, Kunth, has three or four South-African
species with a close almost spikelike panicle and hairy glumes,
allied in many respects to Keleria ; but the inflorescence as well
as the many-nerved glumes bring them nearer to Dactylis.

57. Brizoprrum, Link, as now understood, is specially founded
on the Mediterranean Poa sicula, Jacq., to which are added three
South-African species. The flat broad spikelets with coriaceous
glumes are nearly those of Eragrostis sect. Platystachye ; but the
flowering glumes have seven nerves, and the spikelets are nearly
sessile in a bifarious spike, or especially the lower ones closely
clustered. I have already referred under Distichlis to the Ame-
rican diccious plants for which the name Brizopyrwm. has been
retained by Presl and by Fournier; the true Brizopyra are all
hermaphrodite.

58. Scuzrocuioa, Beauv., is limited to the S. dura, a small
Mediterranean annual well characterized by the inflorescence and
shape of the glumes; the other species, sometimes referred to
Selerochloa, belong chiefly to Cutanda.

59. Briza, Linn., about ten species, of which the typical ones
are chiefly European, though one has now spread over the greater
part of the civilized world ; two sections are entirely American,
tropical or northern. All are characterized by the very concave,
sometimes almost vesicular, glumes enclosing a much smaller broad
flat palea, the grain much flattened from back to front, and some-
times, but not generally, adhering to the palea. The three best
known European species have a very loose panicle with the spikelets
hanging from capillary branches; the Oriental B spicata, Sibth.,
differs in its narrow closer panicle. The American species have been
separated as two distinct genera, which may be. retained as sec-
tions, though with little difference in essential characters. Chas-
colytrum, Desv., has the awnless spikelets of the European species ;
but the ‘panicle, though branched, is much more compact, the

124 MR. G. BENTHAM ON GRAMINES.

spikelets almost sessile. Calotheca, Desv., with a loose spreading
panicle, has broadly scarious awned glumes.

60. Scutsmus, Beauv. (Electra, Panz., Hemisacris, Steud.), has
three or four species, one of them widely spread from the Medi-
terranean region, eastwards to Afghanistan and Arabia and
westwards to the Canary Islands, the others South African. All
are annuals with a narrow panicle, and distinguished by the long
empty glumes quite enclosing the flowering ones.

61. NepHetocutoa, Boiss., limited to the original Oriental
species, is a very elegant little grass with the habit of Aira invo-
lucrata, and is figured in the last part of Hooker’s Icones. The
species added to the genus by Grisebach, for which he was obliged
to alter Boissier’s character, are now restored to Poa.

62. Poa, Linn., is a cosmopolitan genus, chiefly extratropical,
which, after frequent extensions and reductions, has now become
fairly limited to a series of about eighty species. They form a
group natural enough as to the great majority of species, dif-
fering from Eragrostis in their five-nerved flowering glumes, from
Glyceria and Festuca in their glumes keeled from the base; but
here, as elsewhere, there are species apparently intermediate
between these large genera, and several smaller ones are only
separated by characters of little importance. Poa has also been
distinguished from Festuca by the obtuse, always unawned glumes,
and by the non-adherence of the grain to the palea. The former
character is general, but not absolute; several species of Poa
have acute glumes, and in P. lanuginosa, Poir., they bear a fine
point which might almost be termed a very short awn. And as
to the grain, though it is usually free, there are several Chilian or
Australian species and some Asiatic ones where it is adherent to
the palea, as in Festuca, and even in the common Poa pratensis
it is often more or less adherent, whilst there are several true
Festucas where it is quite free.

Most of the widely spread species of this genus are so variable,
that it would require much more research into specific detail
than I can at present bestow upon them to distribute them into
natural groups or sections; and I can only refer to the following
as having been proposed as sections or separate genera :— Pseudo-
poa, proposed by C. Koch as a section of Festuca, includes P. per-
sica, Poir., and two other temperate-Asiatic species, with very
small spikelets and with nearly the habit of Nephelochloa, to which
Grisebach has referred them, but which appear inseparable from

MR. G. BENTHAM ON GRAMINER. 125

Poa notwithstanding the adherence of the grain to the palea.
Leucopoa, Griseb., is the temperate-Asiatic P. albida, Turez.,
with the spikelets rather larger than usual, and somewhat sca-
riose and shining glumes like those of several Chilian species.
Dioicopoa, E. Desv. (Dispar, Doell), is a section proposed for
P. chilensis, Trin., P. lanuginosa, Poir., and their allies, in which
the spikelets are usually, but perhaps not always, diccious. In
the dried state there is very little external difference between
the male and the female panicle ; and there are certainly Chilian
specimens otherwise very near P. chilensis which have herma-
phrodite flowers. Doell also places P. lanuginosa in his herma-
phrodite section, whilst Emile Desvaux describes it as dicecious,
as I have generally found it. The stamens, however, are very
deciduous, and the ovary at the time the stamens are still enclosed
very minute; and it requires careful observation to ascertain the
real absence of the one or the other. It is probable that many
species hitherto supposed to be perfectly hermaphrodite are more
or less polygamous. P. lanuginosa shows, moreover, an approach
to Festuca in the fine though short points to the flowering glumes,
and in the adherence of the grain to the palea. Poidium, Nees,
a Brazilian species, was separated by Nees from Poa on account
of a reduction in the number of flowers to one or two, which Doell
finds to be by no means constant.

63. ConropiuM, Trin., about ten species, from the Levant and
Russian Asia, might perhaps be regarded as a section of Poa.
Ii differs in very little besides the small spikelets containing only
one or two flowers, thus connecting Poa with the Agrostee.
The arctic plant, published by R. Brown as a doubtful Colpodium,
now forms Grisebach’s genus Arctagrostis, included above under
Agrostezx.

64. GrapHepHorumM, Desv., including Scolochloa, Link (Flu-
minta, Fries), and Dupontia, Br., contains seven North-American,
North-European, or North-Asiatic species, very well worked up
and distributed into four sections by Asa Gray. They are all
very near Glyceria, differing chiefly in the hairs surrounding the
flowering glumes, which induced several botanists to refer the
genus to Arundine», though very different in habit and in the
shape and venation of the glumes. The hairs of the spikelets
are, moreover, very variable, shorter than in true Arundinee, very
short in the section Arctophila, and not entirely absent in one or
two species of Glyceria.

LINN, JOURN.— BOTANY, VOL. XIX. L

126 MR. G. BENTHAM ON GRAMINEX.

65. Gryceria, Br., if we include Atropis, Rupr., is a genus of
nearly thirty species spread over the extratropical regions, northern
or southern, both of the New and the Old World. It is very
nearly allied both to Poa and Festuca, differing from the former
in the flowering glumes rounded at the base without any promi-
nent keel, from Festuca in the broader more obtuse glumes, and
the grain usually free from the palea, and from both in the short-
ness of the nerves of the glumes. The habit is somewhat variable,
but as much so in each section as in the whole genus. The two
sections into which it has been divided, often raised to the rank
of separate genera, are:—1. Hydrochloa, Hartm. (Porroteranthe,
Steud., Exydra, Endl., Glyceria proper of many botanists), with
the lodicules connate and truncate or deficient, and the thick grain
only marked on the inner face with a very narrow lined furrow or
quite smooth ; and 2. Atropis, Rupr. (Puceinellia, Parlat.), with
two distinct lodicules and the grain more or less compressed from
front, to back, with a broad furrow or almost flat on the inner
face. But these characters are not constant. The lodicules in
the typical Hydrochloa, G. fluitans, By., though thicker than in
Atropis, and usually connate, are readily separable and occasion-
ally spontaneously free ; in G. aquatica, Sm., they are so short as
to render it difficult to say whether they do or do not cohere, and
in G. nervata, Trin., and in G. pallida, Trin., I can find no trace
of them; in Atropis they are usually, but not always, more deve-
loped and thinner. The shape of the seed and of its furrow seems
to vary from species to species, in so far as I have been able to
procure it well ripened. ;

66. Festuca, Linn., is one of the genera as to whose limits
botanists are the least agreed. With the exception of the exclu-
sion of Cutanda and Brizopyrum, we have followed generally the
arrangement proposed by Cosson and Durieu, which would include
between seventy and eighty species (estimated by some at above
two hundred and thirty), almost cosmopolitan in their geogra-
phical distribution, but most abundant in the northern temperate
regions of the Old World, with not many American and very few
tropical species. They are generally distinguished in the sub-
tribe by the flowering glumes rounded without any prominent
keel at least at the base, and acute or awned at the end, and by
the glabrous grain adhering to the palea. But there are excep-
tions to each of these characters; and some species run very much
into Poa, whilst others are scarcely distinct from Bromus. The

MR. G. BENTHAM ON GRAMINES. 127

following are the most prominent groups established as sections
or proposed by some as independent genera:—1. Vulpia, Gmel.
(Mygalurus, Link), panicle narrow, dense, and usually unilateral,
the outer glumes very unequal, one often minute or almost obso-
lete, the flowering glumes awned, and frequently, but not always,
only one stamen. Ifwe had only the common European species,
this might well have been kept up as a genus; but in the South-
American F. wlocheta, Doell, and F. leptothrix, Trin., the panicle
is loose, as in Hufestuca, and in F. delicatula, Lag., F. setacea,
Parlat., the awn is sometimes very short and the inflorescence
rather that of Hufestuca. The proportions of the outer glumes
vary from species to species. 2. Hufestuwca, comprises the greater
number of the species, with a loose, spreading or narrow panicle,
the outer glumes nearly equal, the flowering ones acute or mu-
cronate, rarely short-awned, and three stamens. Amongst them
Grisebach has distinguished a section Pheochloa, with the ovary
slightly hairy at the top as in Bromus ; but the character is very
variable in F, sylvatica, F. varia, and their allies. Doell has pro-
posed a section Mallopetalum for the Brazilian F. ampliflora,
Doell, apparently the same as the Mexican F. amplissima, Rupr.,
characterized by the lodicules villous at the top. I find these
lodicules fringed with long hairs at the top exactly asin F. fim-
briata, Nees, which Doell places amongst his Festuce legitime
with glabrous lodicules. Helleria, Fourn., is proposed asa genus
for the Mexican Bromus lividus, H. B. K., which Sprengel after-
wards and Kunth himself removed to Festuca, of which it has all
the characters of the awned species. The inflorescence is at first
very like that of some varieties of Bromus tectorum; but as it
advances the spikelets become very much divaricate or reflexed,
giving the plant a peculiar habit. 8. Schedonorus, Beauv. (Am-
phigenes, Janka), comprises F. pratensis, Huds., F. sylvatica,
Host, F. nutans, Host, F. littoralis, Labill., #. Hookeriana and
Ff. scirpoidea, F. Muell., and a few others, tall plants, with loose,
narrow or spreading panicles, awnless glumes, and the grain quite
free from the palea, thus connecting the genus with Poa. Fries
and other Swedish botanists, whilst they rightly referred Beau-
vois’s species of Schedonorus back to Festuca, transferred his generie
name to a very different group, which now forms the sections
Festucoides and Stenobromus of Bromus. 4. Catapodium, Link,
including Micropyrum, Link, differs from Hufestuca in the inflo-
rescence, which is nearly the simple spike of Hordeex ; but the

128 MR. G. BENTHAM ON GRAMINEE.

rhachis is not notched and the spikelets are not quite sessile, the
lower ones often two or three together on a very short branchlet,
not collateral. Nardurus, Reichb., is the F. unilateralis, Schrad.,
differing from the rest of the section in the flowering glumes
mucronate or sbortly awned. Castellia, Tineo, is the F. tubercu-
lata, Coss. and Dur., in which the flowering glumes are minutely
tuberculate and the spike often shortly branched. ardurus
montanus, Boiss., scarcely differs from F. (Vulpia) delicatula,
Lag., and F. cynosuroides, Desf., is also referable rather to Vulpia
than to Catapodium. FF. lolium, Balansa, may really be said to
be intermediate between Festuca (Oatapodium) and Lolium. FF.
unioloides, Kunth, is Brizopyrum siculum. Catapodium fusiforme,
Nees, is Zripogon bromoides, Nees. 5. Scleropoa, Griseb. (Sclero-
chloa, Reichb., not of Beauv.), annuals, often small, with one-
sided panicles, the short rigid branches bearing few almost ses-
sile spikelets, at first erect, then spreading or reflexed, giving
nearly the habit of Cutanda, but the glumes entirely those of
Festuca.

67. Panratuera, Philippi, and 68. Popopsorts, Philippi, are
monotypic genera from the island of Juan Fernandez, both very
near Bromus, but scarcely reducible to it.

69. Bromos, Linn., is a fairly natural genus of about forty
species, generally distributed over the temperate regions of the
northern hemisphere, with a very few tropical or southern spe-
cies. Very near Festuca, with which it is closely connected
through Festuca gigantea, Vill. (Bromus giganteus, Linn.), it differs
generally in the flowering glumes distinctly notched or shortly
two-lobed at the end, with an awn between the notches often not
quite terminal and sometimes slightly twisted, showing an ap-
proach to Avena, and in the grain (always adnate to the palea)
crowned by a little appendage or tuft of short hairs. These cha-
racters are, however, not quite constant ; and the four following
sections into which the genus has been divided run also much into
each other, though some of them are often regarded as separate
genera :—1. Festucoides, Coss. and Dur. (Schenodorus, Griseb.),
consists of B. asper and B. inermis, Linn., B. erectus, Huds., and
their allies, tall perennials, coming nearest to Festuca, with the
awns usually very short or reduced to small points. 2. Steno-
bromus, Griseb. (Anisanthe, C. Koch), mostly annuals, with narrow
spikelets and long-awned glumes. Schedonorus of Fries and other
Swedish botanists, but not of Beauvois, includes both Festucoides

MR. G@. BENTHAM ON GRAMINER. 129

and Stenobromus. 38. Zeobromus, Griseb. (Serrafaleus, Parlat.),
spikelets usually broad and thick, the flowering glumes awned,
and the nerves of all the glumes more numerous than in the pre-
ceding sections. Libertia, Lejeune (Michelaria, Dumort.), is the
B. ardennensis, Kunth, differing from B. (Zeobromus) secalinus in
the lateral lobes or teeth of the flowering glumes produced into
slender points or very short awns. Zriniusa, Steud., is the B.
Danthonia, Trin., very near B. (Zeobromus) macrostachyus, Desf. ;
but most of the flowering glumes, especially the upper ones of
the large spikelets, bear three long recurved awns. 4, Cerato-
chloa, DC. (or Beauv.), three or four American species, extra-
tropical or Andine, with flat spikes not unlike those of Uniola, but
at length often thickened as in Zeobromus, and the flowering
glumes scarcely notched at the end, and the awn very short.
Fournier rightly retains the B. (Ceratochloa) purgans, Linn., in
Bromus (under the name of B. Hooker), but keeps up the genus
Ceratochloa for the original C. unioloides, DC., as having the
lodicules connate. I have examined a number of specimens, both
wild and cultivated, and have always found the lodicules attached
by a broad base and contiguous, but quite free or only exceedingly
shortly cohering at the very base.

70. Bracnyropium, Beauv. (Hemibromus, Steud.), has five or
six European or temperate-Asiatic species, one or two of which
are also in Mexico, Colombia, and tropical and southern Africa.
They closely connect Festuca with Agropyrum ; the spikelets are
those of the former though usually longer, and the simple spicate
inflorescence is that of Agropyrum, except that the rhachis is not
articulate and not at all or scarcely notched, and the spikelets are
not so closely sessile, usually few and distant. Trachynia, Link,
is B. distachyum, Roem. and Schult., which differs from the rest
of the genus as an annual, with only one or two spikelets at the
end of the peduncle.

Tribe XIII. Horp4reEz.

This tribe, one of the most definite of the Poaces, is charac-
terized chiefly by the inflorescence. The spike is always simple,
except in abnormally luxuriant cultivated varieties or monstrosi-
ties, the rhachis notched and often, but not always, articulated,
the spikelets (one- or several-flowered) singly or two or more
collaterally sessile at each notch. The genera, mostly very dis-
tinct, belong to the temperate regions of the New as well as the

130 MR. G. BENTHAM ON GRAMINES.

Old World, chiefly in the northern hemisphere ; and scarcely any
species, except as introduced weeds or escapes from cultivation,
penetrate within the tropics. The twelve genera are readily
ranged in three distinct subtribes, and require but little comment
on the present occasion.

The first subtribe, Zriticee, comprises four genera, in which
the spikelets have three or more, or very rarely only oue or two,
flowers, and are singly sessile at each notch of the rhachis.

1. Lorrum, Linn., is at once distinguished from all others of
the tribe by the position of the flat spikelets with their edge
to the rhachis. Steudel enumerates twenty-two species; most
authors reduce them to three or four, which run much into each
other. De Rouville published at Montpellier a detailed mono-
graph, in which he rejects all the old species and redivides the
genus into three primary and several subordinate races, to which
he gives new characters and new names, doing little but add to
the prevailing confusion. Two genera have been founded on in-
dividual species or forms—Crepaliwm, Schrank, is the L. temulen-
tun, Linn., and Arthrochortus, Lowe, is a Madeiran species or
variety very near to L. rigidum, Gaudin (L. strictum, Parlat.),
and to some varieties of Z. temulentum.

2. Acroprrum, J. Gertn. (Elytrigium, Desv.), contains about
twenty species, formerly regarded as congeners of the cultivated
Wheats, from which they differ much in habit and technically in
the lateral nerves of the flowering glumes connivent at the top
or confluent into the terminal awn. They are well distributed
into two sections:—1l. Ayropyrum proper, mostly perennials,
with the spikelets more or less distant along the common pe-
duncle or rhachis, the outer empty glumes usually very unequal-
sided and not keeled. To this section belong the common -A. re-
pens, A. junceum, A. caninum, and a few others. Regneria,
C. Koch, is, according to Grisebach, a species closely allied to
A. caninum. Anthosachne, Steud.,is the Australasian A. scabrum,
Beauv. (Festuca scabra, Labill.), which, with the closely allied
East-Indian A. semicostatum, Nees, and the Oriental A. longearis-
tatum, Boiss., differs from the commoner species in the denser
spikes and narrower glumes tapering into long awns at length
diverging. A. pectinatum, Beauv., is an Australian species still
further connecting Agropyrum proper with Lremopyrum. 2. Ere-
mopyrum, Ledeb. (Cremopyrum, Schur, perhaps by a clerical
error, Costia, Willk.), mostly annuals, with the spikelets distichous

MR. G. BENTHAM ON GRAMINER. 181

and close together in a short dense spike, the narrow empty
glumes nearly equal-sided and keeled. Two species, A. villosum
(Secale villosum, Linn., Haynaldia, Schur) and A. hordeaceum,
Boiss., form the proposed section Dasypyrum, Coss. and Dur.
(Pseudosecale, Gren. and Godr.), differing slightly from the other
species in the empty glumes rather unequal-sided, and one lateral
nerve on one side of the keel very frequently as prominent as the
keel itself, giving the glume the appearance of being two-keeled.
Heteranthelium, Hochst., from the Levant, is a species very near
A, (Eremopyrum) orientale, with a dense villous spike, and several
of the spikelets, especially near the base and apex of the spike,
often sterile with empty glumes.

3. Secatz, Linn., is now reduced to two species or perhaps
varieties, the cultivated Rye, of which S. montanum, Guss., is
supposed to be the original spontaneous form, and 8. fragile,
Bieb. The genus differs slightly from the section Hremopyrum
of Agropyrum in the dense cylindrical spike, and in the spikelets
usually containing only two flowers.

4. Trrricum, Linn., excluding Agropyrum and including Aigi-
lops, can scarcely reckon more than ten botanical species; the
most prominent character separating them from Agropyrum con-
sists in the shape of the spikelets not so flat, and especially in the
lateral nerves of the flowering glumes not connivent, but remain-
ing parallel or nearly so, and either stopping short of the apex or
produced beyond it into distinct teeth or awns. There are three
rather distinct groups:—1l. The cultivated Wheats, of unknown
origin, in which the flowering glumes are keeled at the end and
sometimes from the base, and terminate in a single awn, the
lateral nerves usually barely reaching to the end of the glume.
2. Crithodium, Link, founded on T. monococewm, Linn. (T. baoti-
cwm, Boiss.), in which the spikelets have only one fertile flower,
and the flowering glume is keeled from the base and ends in
asingle awn. TZ. bicorne, Forsk., with two or even three fertile
flowers and the lateral nerves of the flowering glumes sometimes
produced into short teeth, may be referred to the same section.
3. Aigilops, Linn., above forty published species, which Munro
reduces to seven or eight, differing from the cereal wheats in the
flowering glumes more rounded at the back and not at all keeled,
and in the lateral nerves of the flowering glumes often produced
into long awns, especially in the upper end of the spike. The

132 MR. G. BENTHAM ON GRAMINES.

extreme readiness with which some species hybridize with the
cultivated wheats has given rise to the suggestion, strongly advo-
cated by some, positively rejected by others, that it is in some of
the common species of Avyilops that we must look for the original
of our cereal wheats.

The second subtribe, Lepturee, is characterized by the slender
spikes and the spikelets solitary at the notches, each with only
one or rarely two flowers. We refer to it five genera, placed by
Kunth and some others in Rottboelliew, from which they differ in
the outer empty glumes, when present, persisting below the arti-
culation of the rhachilla.

5. Lerrvrvs, Br., including Pholiurus, Trin., has six species,
five of them with the ordinary geographical range of the tribe,
the sixth, Z. repens, Br., exclusively Australasian or South Pacific
and maritime. They are distinguished in the subtribe by the
rigid outer empty glumes, one or two in number, much longer
than the hyaline flowering glume, thus showing the nearest
approach to Rottboelliew. They differ from each other sufficiently
to have been referred by different botanists to different genera.
L. cylindricus, Trin. (ZL. subulatus, Kunth), included by Link in
Ophiurus, by Reichenbach in Monerma, has one outer empty glume
and one flower with no empty glume above it. The Australasian
LL. repens, a much larger plant than any of the others, has one
outer empty glume, one flower, and above it a glume either
empty or enclosing a palea, but no flower. L. persica, L. ineur-
vata, and L. filiformis, Trin., have two lower empty glumes, one
flower, and no empty glume above it. JZ. pannonicus, Kunth,
forming Trinius’s genus Pholiurus, and referred by T. Nees to
Ophiurus, has two outer empty glumes and two perfect flowers.

6. Psizurvs, Trin. (Monerma, Beauv., partly, Asprella, Host

but not of Willd.), is a single annual, near Lepturus, but with.

only one minute empty glume, a single narrow and awned flower-
ing glume, and only one stamen in the flower.

7. Nagpvus, Linn., is a single well-known small perennial, the
position of which in the system is rather puzzling. The spikelet
has only one flower without any empty glumes below it or pro-
longation of the rhachilla above it, which might have decided its
relationship either to Panicacez or to Poacer, and its long simple
style might indicate an affinity to some Panicezx or to Seslerie ;
but on the whole it seems nearest allied to the Leptures, a sup-
position which might be confirmed, if we regard the rather pro-

MR. G. BENTHAM ON GRAMINED. 133

minent lower margin of the notches of the rhachis as a rudimen-
tary glume.

8. Kratrxra, Coss. and Dur., is a single Algerine species
unknown to me, but well described, and evidently rightly placed
in the present group. 9. Oropgriom, Trin., is a dwarf East-
Indian species remarkable for the cylindrical spike, with per-
fectly immersed spikelets as in some Rottboellias and Ophiurus,
but with the outer persistent glumes of Hordeew.

The third subtribe Zlymee comprises three genera, in which the
spikelets are two or more, collaterally sessile at each notch of the
spike, or the lateral ones very shortly stipitate.

10. Horpeum, Linn., was restricted by Beauvois to the common
cultivated barley, H. vulgare, Linn., which in a great variety of
forms is of very ancient cultivation, and whose indigenous origin
is no more known than that of our wheats. Amongst these
forms the Hast-Indian H. egiceras, Royle, has been proposed by
E. Meyer as a genus under the name of Critho; but it cannot be
otherwise considered than asa luxuriant monstrosity. The really
spontaneous species of Hordewm amount to about twelve, distin-
guished from Elymus by the single flower in each spikelet, and
distributed into three sections:—1. Zeocriton, Beauv. (Critesion,
Rafin.), for the H. murinwm, H. bulbosum, H. jubatum, Linn., and
some others, in which the central spikelet alone of each three is
fertile, the lateral ones sterile or reduced to empty glumes;
2. Crithopsis, Jaub. and Spach (sect. Medusather of Elymus,
Griseb.), for the H. erinitwm, Desf., and its allies, with two per-
fect spikelets at each notch, the intermediate one deficient or
rarely represented by one or two empty glumes; 3. Cuviera,
Koel., for the H. sylvaticum, Huds. (Elymus europeus, Linn.),
with three collateral spikelets.

11. Exyuvs, Linn., as now generally limited, comprises about
twenty species, distributed into three sections, all distinguished
from Hordewm in having two or more flowers to each spikelet :—
1. Sttanion, Rafin. (Polyantheriz, Nees), for the North-American
E. Sitanion, Schult., with the flowering glumes usually three-
awned; 2. Clinelyna, Griseb., with the spikelets usually two only
to each notched and the flowering glumes with one long awn;
and 3. Psammelyna, Griseb., tall rigid species with often more
than two spikelets to each notch, and the flowering glumes un-
awned or with only very short awnlike points.

LINN. JOURN.—BOTANY, VOL. XIX. M

184 MR. G, BENTHAM ON GRAMINES.

12. AspPRELLa, Willd. (Hystriz, Moench, Gymnostichum,
Schreb.), has three species, of which two are North-American,
the third from New Zealand. There are two or three collateral
spikelets to each notch as in the preceding genera; but the outer
empty glumes are entirely deficient, except sometimes one or two
slender ones to the lower spikelets of the spike. Willdenow’s
name has the priority over Schreber’s ; for although the Beschrei-
bungen of the latter author bears the date 1769 on the titlepage,
the third part, in which the poet genus was proposed, was
only issued in 1810.

Tribe XIV. BamsBuseEz.

The Bamboos have been so admirably monographed by Munro
in the twenty-sixth volume of the Linnean ‘ Transactions,’ that
I have very few notes to make on the present occasion. Since
the appearance of that memoir, Balansa has published a New-
Caledonian Bamboo forming the distinct genus Greslania; a fur-
ther acquaintance with Thamnocalamus has induced its reunion
with Arundinaria; and, on the other hand, Merostachys capitata,
Hook., is so very different in inflorescence from the rest of the
genus, that I have proposed to separate it under the name of
Achroostachys. Ihave also proposed as a new genus Melocala-
mus, the Pseudostachyum compactiflorum, Kurz, published since
Munro’s monograph. There is also much confusion in the generic
term Beesha, which, though used by Rheede for a Peninsular
species of Bamboo, was first characterized by Kunth chiefly from
the more eastern Bambusa baccifera, Roxb., now Melocanna bam-
busoides. He did indeed also include the Peninsular and Ceylon
species; but that was first properly characterized as a separate
genus by Thwaites, under the name of Ochlandra, which it seems
advisable to adopt, though the genus may include Rheede’s Beesha,
a name which it seems best to consider only in the specific sense
first given to it.

DE. W. R. M‘NAB ON ARCTIC DRIFT WOODS. 135

Report on the Arctic Drift Woods collected by Captain Feilden
and Mr. Hart in 1875 and 1876. By W. R. M‘Nas, M.D.,
F.LS., Professor of Botany, Roy. Coll. of Science, Dublin.

{Read November 3, 1881.]

In the following Report I have endeavoured to detail the results
of my examination of the drift woods brought from the Arctic
regions by the naturalists attached to the recent Arctic expedition
under Captain Nares. There are thirteen specimens of drift
wood and one specimen of bark, collected in different localities by
Captain Feilden and Mr. Hart; and these were placed in my
hands for examination by Professor Oliver, F.R.S. The speci-
mens of wood are all completely devoid of bark; hence it was im-
possible to distinguish the genus to which some of the Coniferous
woods belong, as, for example, Picea and Larix, the genera to
which most of the woods may be referred. In general the
woods were well preserved and in good condition, except on the
very surface ; hence there was little difficulty in obtaining proper
sections for microscopic examination. The woods were all cut
in three directions, as is usual in examining dicotyledonous and
coniferous woods; and the sections were viewed both dry and
when mounted in Canada balsam. Careful comparison of the
drift woods with sections of named woods has not enabled me to
identify the species in any case ; hence the whole of the results
obtained must be considered unsatisfactory. In the following
list the specimens are lettered A to O, the letters &c. correspond-
ing to the labels on the slides of the preparations accompanying
the Report*.

1. Pinus sp. (One species.)

Two portions of wood are referable to Pinus. These I have
indicated by the letters A and B. Both the pieces in the collection
are marked as from the same locality, viz. “ Head of Discovery Bay,
April 1876,” but one of them (B) as having been “ 100 yards from
the water, embedded in sand.” The woods are quite similar in
outward appearance, and are portions of comparatively large

* [have also examined a small collection of drift woods made by Staff-Sur-
geon Edward L. Moss, M.D., R.N., Surgeon of the ‘Alert.’ His specimens are
similar to those obtained by the naturalists of the Expedition, with one excep-
tion, viz. a portion of a stem of a species of Juniperus. The stem must have
been of some size, and belongs apparently to a North-American species.

DR. W. B. M‘NAB ON THE ARCTIO DRIFT WOODS

186

and I am inclined to consider

them portions of the same species of Pinus, if not parts of the

same tree.

The central portions of the stem are well

preserved ; but the outer part is soft and partially destroyed.

A has well-developed annual rings, some of the rings near the
periphery of the stem being, however, small and feebly developed,

thus indicating a deficiency in growth. The wood of B pre-

sents the same general characters ;

trunks or branches.

|

parallels 78°-83° N. lat., with the localities generally where the drift

Sxetcu Map showing the route of the late Arctic Expedition between the
woods were obtained.

COLLECTED BY CAPT, FEILDEN AND MR. HART. 187

2. ABIES sp. (One species.)

One piece of wood of small size; but probably a portion of a
larger stem cut and rounded by man, belongs to the genus Abies
(C). By comparison with other sections, it seems to come very
close to Abies pectinata. The specimen is marked “ Discovery
Bay, 20 feet above sea-level, Aug. 1875, Captain Feilden.” The
annual rings of wood are large and well developed.

8. Prcea or Larix. (Three species.)

Seven pieces of drift wood are to be referred to one or other of
these genera; but, owing to the absence of the bark, it is impos-
sible to decide definitely. One of the specimens (D) comes very
near Larix, and differs from all the other woods in the collection.

P Larix sp. This is the specimen D, marked “ Dumb-bell
Harbour.” The stem has been large, and is well preserved, and,
by comparison with named sections of Larix, seems to come very
near L. europea.

? Proza sp. Twospecimens seem to belong to one species, viz.
E, “Label incomplete. Upon Floe. Sept. 12, 1875,” and F,
“On Floe, lat. 82° 30' N. Capt. Feilden.” These are portions
of well-preserved woods, white and firm, and having the same
microscopical characters. They are probably not portions of the
same stem, as I believe E is almost certain to be from Mr.
Hart’s collection, while F is Captain Feilden’s. The annual
rings are well developed in both specimens.

?Proma sp. Four specimens. G, H, J, K. All similar in
microscopic character, and belonging either to Picea or Laria.

(G) “1 mile inland and 150 feet elevation at ‘Alert’ winter-
quarters, Feb. 1876. Captain Feilden.” This piece of stem has
well-developed annual rings.

(H) “Drift wood. Bottom of Musk-ox Fjord. Sept. 16,
1875.” 85:5 inches long, 16 in circumference. Portion of a
large stem with well-developed annual rings.

(J) “No. 1. No locality nor date.” A small piece of very
much waterworn drift wood with well-developed annual rings,
and probably a portion of a large stem.

(K) “No. 1009. No locality.” Small portions of a large stem
in a good state of preservation, and having woll-developed annual
rings.

Proza sp. Bark only. The specimen (L) is from pieces of
bark evidently of a Picea, and marked “ On floe in Dumb-bell Bay,

188 DR. W. B. M‘NAB ON ARCTIC DRIFT WOODS.

Sept. 1875. Found by Commander Markham.” It is not impro-
bable that this may be the bark of the commonest drift wood, and
may therefore help to identify the genus of the six specimens of
wood just described.

4, Taxus sp. (One species.)

A single sample of very much waterworn pieces of drift wood,
which have probably been embedded in mud, is referable to the
genus Taxus, having the spiral markings clearly shown in the
wood prosenchymatous cells. Itis marked M, “Out of cloth
bag ; uo locality.” The annual rings are extremely imperfect and
very numerous.

5. Popunus sp. (One species.)

Two pieces of drift wood are to be referred to Populus (near
tremula), and are interesting as being the only species of dicotyle-
donous wood in the collection. One of the specimens (N) is
marked “ Drift wood. Musk-ox Bay. Sept. 1875 ;” and the other
(O), “East Cary Island. Capt. Feilden.”” The former isa por-
tion of a large stem, and is in an excellent state of preservation,
while the latter is equally well preserved, but is only a part of a
small branch. In both of the woods the annual rings are well
developed.

Of the 14 specimens submitted to me for examination, 18 are
samples of wood, and 1 is of bark alone without any trace of wood,
the bark being evidently coniferous and to be referred to the genus
Picea. Ofthe 13 woods, 11 are coniferous, and only 2 dicotyledo-
nous, both belonging tothe same genus, Populus, and to the same
species.

The 11 coniferous woods belong to four, or perhaps five, genera,
there being 1 species of Pinus, 1 of Abies, 1 of Larix ?, 2 of Pinus
or Larix, and 1 of Taxus. Of Pinus there are 2 specimens, of
Abies 1, Larix or Picea'7, and Taxus 1—the commonest form being
some kind of Picea, probably an American Spruce.

I have not been able to identify the species, but, from careful
comparison of specimens, am inclined to think that most of them
are North-American; and as the annual rings are usually very well
developed, the trees must have grown in the more northern tempe-
rate latitudes.

. ‘NOLNIVIS ‘L “H
suonpny {

' ‘d8Ia0 INVUA
GSaeud “f SHTUVHO

“TSST ‘FT Lepr

004.100, PUNO} pus peuruTexe Useq aavy s}UNODIe SuroFe10F oy,

wounsnaay, ‘KAMANO MOTMACTAA

L OL SLEeF
oro seg oo asia qotnente “vstoqueg oy} Jo spury omp Ur eouDTeE
0 0 OFLC * quomysoauy
6 6 109IF
& 81 994 ———_ F
BG Gowe ccs sols adiclcaroauion Soyer
PYM Sh GO 5s So aaa Suryug
— suonvorqug s£yo100g jo sosuedxny
TL I 8 AJQuoyeyg pues Surpurqyoog
G9 QZ | teeter tenets stereeee eeeee tee eeeneee gayug SNOdUET[EOSTPL
$ 8 FOL Pore eee eee ey Heed e ene e wees postyind syoog
8 GL 9g cute (e8eysog pue voy, Surpnpour) sesuedsg AWWog
UF ee eee UWOISSIUINIO() 8 AOJOATJO) pu ‘TorwsUsg ‘selreleg
IL 6 T¢ adel aiadiisbiasjeis Meuse taistaa dsinws oy eco esee eats BeH pur E[LOQ
9 16 adtQeinieavainnies saC ogi Vented ane RTT omymumyg pur saredoy
6 SL GL ces Vteeenseseneer cores gouBINSUT PUR SOXVT,
Pe F

L OL SLOorF

“Weyjueg “IP, Worf UoryeUOg yewedg

SUOTIVOTTgN, spreMo} suOTYeUO(T

mdopeyen Areaquy

“sTeuanoe

** suOOUsMEL],

—: stoMVoTTQng 84421009 jo soTeg

saved quasetd of

‘+ gavak azorad a0f

—0 0 € J ONTP ont
" Q & _ GF Jo suoynquyuey yenuTy

jaa Himes saneneoie Herarans Gormaniie suomsoduoy

sea UOIssrpy

SpUN,T poyseaT] UO ysexoqyUT

* OSSI ‘AePL 487 ‘saequeg 42 coueleg

oooo co nmeooo

“ISSL ‘aIFs AvpL ‘Aepsony,

‘Buyoop, Aresioatuny 04} 48 poyWosorg

‘ISS ‘Tady 430g Surpuo rvok oy4 10} yUNODOw sNaNASYEUY, THY,

‘ALGIOOS NVANNIT

EVENING MEETINGS OF THE LINNEAN SOCIETY.
To BE HELD AT Buruinaton House, Session 1881-82.
On Thursday Evenings, as undermentioned, the Chair being taken at 8 pnt.

1882. January 19 1882. March 16 1882. May 4
February 2 April 6 June 1
1 » 20 ag

March 2

The Anniversary Meeting takes place on Werpnuspay,
24th May, 1882, at 3 p.m.

Memoranda concerning Communications read before the Society.

The Council desire it to be understood that Authors are alone respon-
sible for the facts and opinions contained in their respective papers.

It is to be noted that the sequence of the papers as printed in the
Society's Journals do not, in all cases, absolutely follow date of reading.
Some communications require reconsideration of Council; and others de-
pend on exigencies and convenience in printing, illustrations, &c., which
may delay or expedite their publication. As far as possible, precedence is
given to papers in order of reading, especially when not very long or
complex in kind. With contributions of a lengthened or technical cha-
racter, or where the author cannot be present at the reading of his paper,
as in the instance of Fellows resident ane the business of the Meeting
and interest of the writer will be greatly facilitated if an abstract for
reading be sent with the manuscript. sll drawings for illustrations should
be accompanied by full descriptions. °

MSS. &c. may be addressed to the President, the Secretaries, or Librarian, at
the Society’s Apartments, Burlington House, Piccadilly, London, W.

Specimens intended for exhibition, or diagrams, maps, and objects intended
to illustrate papers to be read, should, if convenient, be sent to the Society’s
Rooms previous to the hour of Meeting, and accompanied with any memoranda

concerning them.

THE LIBRARY AND READING-ROOM.

The Library is open to the Fellows and their friends daily, between 10 a.m.
and 4 p.m., and on Meeting nights at 7 p.m.; and the Reading-room adjoining
from 10 a.m. till 6 p.m, except when Council is sitting and on Saturdays only
till 4 p.m. With certain restrictions, Fellows are allowed to borrow Books.

LIBRARY COMMITTEE.

This consists of nine Fellows (three of whom retire annually) and of the four
officers ex officio, in all thirteen members. The former are elected annually
by the Council in June, and serve till the succeeding Anniversary. The Com-
mittee meet at 4.p.m., usually once a month. The Members for 1881-82, in

addition to the officers, are :—

John Ball, M.A., F.R.S., M.R.LA. | Albert C. L.G. Gimther, M.A., M.D.,
George Busk, F.R.S., F.G.S. F.RB.S. :

William Carruthers, F.R.S., F.G.S. Edward Morell Holmes.

Prof. P. M. Duncan, F.R.S., F.G-S. Howard Saunders, F.Z.S,

W. T. Thiselton Dyer, M.A., F.R.S. Henry T. Stainton, F.R.S., F.G.S8.

A Book for insertion of Recommendations of Volumes to be added to the
Library lies on the table in the Society’s Rooms at the disposal of the Fellows.

Notz.—The Charter and Bye-Laws of the Society, as amended up to
the 21st April, 1881, have now been reprinted ; and any Fellow can have
a copy of the same on application.